Reservoir simulators are effectively used in predicting the production performances of oil and gas reservoirs within a good level of accuracy. However, during initial stages of exploitation, our knowledge on most of the reservoir properties and production parameters involve a good degree of uncertainty. In such cases, expert systems can be effective as a screening tool by mimicking the performance of a reservoir simulator at a lower cost and reduced personnel and machine time. The expert system described in this paper is a tool which can be used to predict the performance of a coalbed methane reservoir similar to its numerical counterparts.
The proposed expert system has been trained with the help of an extensive data base and has the capability of providing gas and water production profiles for a period of about ten years for a given coalbed methane reservoir. Other outputs include cumulative gas and water production, peak gas flow rate expected, time to achieve a peak rate and abandonment time. This study also involves the development of an inverse expert system which has the capability of identifying the optimum production strategies for a desired production scenario from a coalbed reservoir. Conventional reservoir simulators cannot suggest an optimized design strategy that can help achieve a certain desired production performance from a coal seam unless a large number of scenarios are studied. The inverse model can tackle this optimization problem quite effectively.
During initial stages of exploitation of oil and gas reservoirs, there is always a good level of uncertainty in determination of reservoir parameters. In such cases, there is a need to develop a screening tool that can predict production performance from reservoirs for several possible combinations of reservoir properties and design parameters. By training an artificial neural network with various physically possible combinations of these parameters, it is possible to come up with a simple and cost effective tool that can mimic production from coalbed methane reservoirs for a certain time of production.
Most of the parameters that play a significant role in impacting gas production are included as inputs in the process of development of the neural network model. Inputs to the model can be grouped under two categories - reservoir properties and production design parameters. Neuro simulation methodologies applied in this work involve coupling of hard-computing and soft-computing procedures. Data samples required to train, validate and test the tool box are obtained using PSU COALCOMP, a three dimensional compositional CBM reservoir simulator developed by Manik and Ertekin in 2002.
Artificial neural networks are computational models that are developed on the principle of functioning of the biological nervous system. They work on the hypothesis that a desired level of intelligence is achieved by means of interaction of a number of simple processing units called nodes (Jeirani, Z., et al 2005). Mathematically, in an ANN, a set of weights are found that generates outputs when the network is presented with an input set. This process is called learning and by altering the weights of links between nodes, the learning abilities of such neural networks are improved (Mohaghegh et al, 1994). The most important characteristic of neural networks is their adaptability. By having them exposed to sufficient examples, they can learn by adjusting the links and connections between the neurons.