Good petroleum reservoir management focuses on maximizing profitability, production and ultimate reserves recovery by integrating technical, commercial and risk management in a dynamic and uncertain environment. Constraints (technological, physical, financial, geopolitical, environmental, safety, corporate and human) limit the decision options available.

Figure 1. Traditional framework of oil and gas data sources, technical, economic and risk interpretation models and decision-making tools. (Images courtesy of the University of Wyoming)
Deploying new and better technology reduces constraints and provides more and better quality information about the reservoir that ultimately should improve reservoir performance.
Intelligent or “smart” technologies deployed in well bores aim to provide better remote well performance monitoring and identify early when reservoir interventions are required. Optimized well performance means less lost production through downtime or inefficient operating conditions and more profitability from production operations. The economic viability of such technologies requires careful cost vs. benefit and risk analysis (Figure 1).

In remote field developments they can provide extremely cost-effective solutions, particularly where the information they provide can be accessed and interpreted remotely from the well site. However, such tools have yet to reach their full potential as the wealth of data they provide is not processed and acted upon quickly enough. This is set to change with the development of dynamic real-time optimization reservoir models that should complement the more cumbersome traditional full-field multiphase reservoir simulation.

To achieve maximum impact, smart technologies should be deployed and the information they provide integrated with standard reservoir and production management tools, databases and models to contribute information to guide (and be guided by) real time reservoir optimization models.

Optimized reservoir management

Optimization of oil and gas assets, either for profitability, cost, production or reserves recovered usually combines mathematical models, field data and experience to influence investment or methodology decisions. For example, a reservoir simulation model fed with up-to-date reservoir information from well testing and downhole tools provides multiple future production and reserves recovery scenarios based on existing and carefully placed future wells (production and injection). Decision makers expect reservoir managers to select the best options from the multiple simulation scenarios. Optimization software can help reservoir managers to do this and monitor performance of selected options as new production and reservoir data is collected.

If a reservoir simulation model is not frequently updated with new data and new history matches conducted, the “optimized” solutions rapidly become irrelevant and may then be put aside to return to traditional tried and tested decline-curve and water-cut analysis, which provide the reservoir managers with an understanding of what is actually happening in active wells but offers little in terms of optimization solutions.

Reservoir models and mathematical optimization routines are not one-off exercises but need
Figure 2. Production and reservoir dynamic real-time optimization model methodologies.
to be repeated frequently and acted upon when new unplanned conditions prevail and production data becomes available. One problem is that updating the reservoir simulation history-match can be a laborious task unless data is provided systematically and in an appropriate format. Unwieldy and slow integration of data into many reservoir simulation models to ensure the adjustment of model metrics to match observed production, water-cut and reservoir-pressure history commonly make the reservoir simulation unsuitable for accurate short-term predictions and decisions.

Reservoir models that can make reliable short-term predictions are essential for production-related optimal decision making. The increased availability of real-time data in the field derived from intelligent completions can provide the necessary information to feed short-term, data-driven reservoir optimization models. Indeed, permanently instrumented wells that can be remotely actuated and interrogated facilitate reservoir optimization performed in real time based on easy-to-manipulate models updated by regular feedback from active wells.

Real-time optimization

Real-time optimization (RTO) is a method frequently used in the downstream industry for complete or partial automation of the process of finding good (optimal) control settings. By continuously collecting data from a process plant, the data are analyzed and optimal plant control settings are found. These settings are then either implemented directly (closed-loop) in the plant or they get presented to an operator (open-loop) for interpretation and a decision to implement or not. The main aim of RTO is to improve utilization of the capacity of a production plant to get higher throughput and improve efficiency and profitability. The model is then continuously updated with new plant measurements and the best fit of the actual input-output behavior of the plant is repeatedly recalculated and an optimization control directive is issued.

This RTO approach can be adapted to optimize performance from an oil or gas reservoir, well production and field process facility complex. A general RTO system used in downstream plants consists of a five-step procedure:
• Data validation;
• Model updating;
• Model-based optimization;
• Optimizer command evaluation; and
• Decision to adopt or reject optimized solution.

Self-learning models involve processes by which a system uses its own past operating data to progressively further develop and refine evolving algorithms as each new batch of new data becomes available.

Integrating field data for continuous learning of key reservoir features based on simplified hybrid models and multilevel optimization is more suitable for real-time operations than full field-wide optimization. Multilevel decision-making has also been developed and extensively tested in the refining and petrochemical industries and can be adapted to reservoir optimization.

Self-learning models

Hybrid, self-learning reservoir models are being developed and exploited when data are scant, as is often the case in practice early on in reservoir development. These can balance the accuracy of data fitting with the model’s predictive ability by appropriate selection of model algorithms. Hybrid models may employ a first-principles structure along with empirical constitutive equations (e.g., Darcy’s law, ideal gas law, pressure-drop relationships) and rely on incoming data to identify and regularly update values of many of the algorithms’ parameters. Because of this, hybrid models are often easier to develop and manipulate than raw first-principle simulation models, while maintaining model fidelity outside the range of the data used for model parameter identification.

The desired model structure is a self-learning adaptive scheme that optimizes multiphase fluid migration in compartmentalized reservoirs, while integrating downhole completions, wellhead restrictions and business constraints. It should continuously optimize reservoir performance while satisfying surface and sub-surface constraints.
Various proposed RTO methodologies involve dynamic models for short-term forward planning. Highly variable and hard-to-process feed data arriving from the wells equipped with intelligent completions make it hard to adapt steady state RTO solutions from the downstream process industry. It is the dynamic nature of the model types described and the multilevel optimization that allows them to process, learn from and act upon uncertain and varying feed (Figure 2).

By continuous processing of data being delivered from remote sites, RTO systems can identify and respond quickly when plant and well go offline or move outside normal operating conditions. This could have huge benefits in potentially preventing hazardous outcomes and improving safety and environmental management. These benefits and the reservoir management benefits all progressively reduce uncertainty, which should ultimately lead to reduced operating costs.