| Title | Optimal Well Design for Stimulation of Geothermal Wells with Radial Jet Drilling |
|---|---|
| Authors | Olwijn LEEUWENBURGH, Elisabeth PETERS, Diederik TROOST, Holger CREMER |
| Year | 2020 |
| Conference | World Geothermal Congress |
| Keywords | optimization, radial jet drilling, geothermal energy, uncertainty quantification, well design |
| Abstract | Geothermal wells can be stimulated to increase productivity or injectivity by drilling multiple, small-diameter laterals using a technique called Radial Jet Drilling (RJD). One would like to design the placement of these laterals such that an optimal connectivity with the reservoir is achieved. Since there is large flexibility in the design in terms of number of laterals, the along-well kick off depth, and the orientation and length of laterals, as well as complex interaction between the dynamic effects of changes in the design, determining the optimal design may be fairly challenging. This complexity is further increased if the reservoir is heterogeneous or if the distribution of rock properties and the trajectories of laterals are uncertain. Here we will investigate the sensitivity of the geothermal system performance to the design parameters for an example case involving a pair of geothermal wells positioned in a reservoir with properties typical for the Dutch setting. We will investigate the sensitivity directly by simulating different scenarios for the design parameters, and we subsequently explore the potential to use numerical optimization techniques to find optimal values for these parameters. Both (local) gradient-based and (global) evolution strategies are considered and we comment on advantages and disadvantages of both in the context of the design of geothermal wells. We find that both approaches are able to deliver improved designs in different settings, but that stochastic gradient methods may be best suited for design optimization under geological uncertainty. |