OPTIMIZATION FOR DRILLING SCHEDULE OF WELLS IN THE DEVELOPMENT OF HEAVY OIL RESERVOIRS

During the development stage of a petroleum field, one important decision is to define the schedule for drilling the wells. Several general rules were listed for light and heavy oils. However, these rules are not always applicable and it may be important to use simulation models to test and choose the schedule. This paper consists of the development, implementation, and application of two different algorithms for optimization of wells drilling schedule. The first algorithm seeks, for each period, which well brings the best economic output. Once this well is selected, the second period of time is tested considering the remaining wells and this procedure is repeated until the last well. The second procedure is based on the reduction of search space where random schedules are generated and the best results maintained for the subsequent generation of scenarios, only allowing the wells to be drilled in the period that produced the best values of the objective function in the previous step. This procedure is repeated until each well converges to the period that results in the best economic return. Both algorithms were tested in two synthetic fields, based on the characteristics of offshore heavy oil and high average permeability reservoirs. To generate a benchmark for the solutions, a large amount of random schedules were tested and a normal distribution for net present values was generated. Both algorithms can be applied in any type of reservoirs, resulting in a very time consuming process in the cases where simulation time is very high. The results from both algorithms lead to net present values higher than at least 95% of the values from random schedules. For both cases, economic results were significantly better than those found for selecting strategy using wells ranking based in economic indicators, which is a common procedure. Both algorithms are also easy to implement and they can be inserted in a cycle of automated or assisted optimization process.