FMS

Forecast Management System

The tool deployment has reduced the time spent in the forecast process by 50% and improved the average forecast accuracy by 40% compared to conventional systems

Production forecasting for oil and gas is one of the most complex and critical tasks in reservoir engineering. It directly impacts field development plans and is essential for reserve estimation and economic evaluation. The process of forecasting undergoes many cycles with different time frames and objectives. It relies heavily on large amounts of historical data previously stored in the form of excel sheets and distributed among many stakeholders. FMS was initially developed to overcome all the challenges associated with the lack of reliable source of data. Automated DCA has been as well implemented to produce different estimates based on three

commonly used models, namely, exponential, hyperbolic and harmonic. Reservoir engineers can select the adequate and relevant historical data by varying the initial start date for the well under consideration. Allowing multiple scenarios based on various models and start date, gives the reservoir engineers the opportunity to select the most accurate forecast. Various ML time series prediction algorithms are also integrated and resulted in 40% improvement in the forecast accuracy over test wells compared to the conventional DCA models.

WELL MONTHLY PRODUCTION FORECAST OF OIL

ADVANTAGES OF ZENN FORECASTING TOOLS

TECHNOLOGY

TAILORED SOLUTIONS COMBINING DATA SCIENCE AND DOMAIN EXPERTISE

DATA

MAXIMIZE THE VALUE OF AVAILABLE DATA BY OPTIMIZING THE OPERATIONS TO INCREASE EFFECIENCY AND PROFITABILIY

KNOWLEDGE

ENGAGEMENT OF YOUNG TALENTS IN THE DEVELOPMENT OF SOLUTIONS

The development of FMS, allowed the team to research, implement and test different neural network architectures. Practical aspects, such as the setting of values for hyperparameters and the choice of the most suitable frameworks, for the successful application of deep learning to time series prediction, were thoroughly investigated.