Veranstaltungsprogramm

Reservoir simulation in the oil & gas industry has been developing over several decades. Fundamental equations and approaches naturally have an even longer history than the software for numerical simulation, which became practically usable in the 70s. Since then the capabilities of numerical simulators have been further pushed along with the increasing computing power. Starting with so called blackoil (oil,gas,water) simulators, the capabilities were extended to cover e.g. compositional description of hydrocarbons, thermal effects of steamflooding, chemical processes for polymer and surfactant injection and more. During the recent two decades the integrated modelling of geophysical, geological, petrophysical, reservoir and production engineering parameters and processes as well as the modelling of the mechanical behaviour of rocks especially in the vicinity of wells became more and more important. The ever increasing resolution of our digital models as well is aiming to improve our predictions. In more recent years such supporting methods like assisted history matching with multiple scenarios, uncertainty evaluations and artificial intelligence became available and are further developed. Most of these developments are fully available with commercial simulators at your fingertips with a mouse click to solve your problem.

Well - not quite. Still the database of measured data typically has a comparable size, resolution and precision like decades ago. Although there has been significant development in the industry to improve the quality and amount of available measured data, still the ratio of the known over the unknown stays at a very similar level. And the increasing resolution of our models mentioned above does not necessarily do the trick.

In principle the non-uniqueness of a solution of an inverse problem under uncertainty prevails. So, how to deal with the vast amount of at least partially uncertain or unknown parameters to be combined in a coupled model comprising hydro, thermal and mechanical processes (not to forget the underlying geological model) in order to match the history and predict the future? Two main approaches how to deal with uncertainty used in the oil & gas industry are presented.

Testing or enhancing the productivity of geothermal prospects in the Bavarian Molasse Basin is related to very low or to no induced seismicity. This is due to the geomechanically favorable orientation of the predominantly E-W striking fault planes relative to the N-S oriented major horizontal stress, resulting in high normal and low shear stresses acting on the fault planes. As observed in some geothermal projects, these stable fault planes may however only yield insufficient fluid productivity. One way to overcome this is to flush proppants into the N-S striking joints, creating potential fluid conduits that drain the nearby fault zones. This approach is currently being tested at the geothermal site in Geretsried within the research project ZokrateS supported by the German Federal Ministry of Exconomics (BMWi) and conducted by Enex Geothermieprojekt Nord GmbH & Co KG, G.E.O.S. Ingenieurgesellschaft mbH, Geothermie Neubrandenburg GmbH, the Leibnitz Institute for Applied Geophysics and the Ruhr- University Bochum. The project aims at showing the validity of the approach.
To study the risk of induced seismicity during hydraulic optimisation, a numerical thermo-hydromechanical (THM) model was run by means of Finite Element Modelling. This was done using the commercially available software package COMSOL Multiphysics. The simulations show that pore pressure perturbations do not affect the stability of the fracture network. Temperature effects are confined to a radius of only a few meters in the short period of operation, thus not affecting the stability of the fracture network. The model has been calibrated against data from the preceding research project Dolomitkluft and shows that the planned injection pressure and rate are unlikely to cause induced seismicity. Furthermore, the fully coupled THM simulation predict a maximum opening of the fracture network of 1.5 mm connecting the fracture network with adjacent faults. This could significantly improve the productivity of the geothermal system.

In the framework of the joint research project ZoKrateS a fully coupled thermo-hydro-mechanical 3D reservoir model was developed to simulate the thermoporoelastic behavior of a naturally fractured stress-sensitive geothermal reservoir in ultra-deep Upper Jurassic carbonates with ultra-low matrix porosity and permeability. Based on measurements and interpretations of a 3D seismic survey in the license field Wolfratshausen ca. 40 km south of Munich, logging, mud losses, zones of joint calcites and hydraulic tests in the 5700 m MD long borehole GEN-1ST-A1 near Geretsried, hydraulically conductive faults and fractures were accurately implemented in a 3D model. For dynamic finite-element analyses, SKUA-GOCAD^TM and COMSOL Multiphysics^® software were used to build a watertight and self-consistent volumetric, unstructured finite-element grid, optimally refined at structural elements to accurately represent the complex geometry and to capture high hydraulic gradients. Hydraulic data obtained in the previous project Dolomitkluft together with the reevaluation of pressure transient analyses of a hydraulic test conducted in the highly deviated Sidetrack, completed with a coiled tubing pipe, a liner and a more than 1 km open hole section in tight carbonates, were used for the hydraulic calibration. Poroelastic effects of the reservoir are assumed to play an important role based on the observations made during drilling operations and hydraulic tests, which suggest that matrix reservoir deforms and fractures close up during increasing effective stress, leading to changing fracture permeability and production performance. Total mud loss occurred while drilling the borehole through the damage zone of the northern synthetic Gartenberg-North fault of the E-W- striking Gartenberg fault zone, a half-graben with two border faults and one minor branch fault in Upper Jurassic carbonates of the Southern South German Molasse Basin. Later productivity tests only resulted in a productivity of ca. 5 l/s compared to ca. 20 l/s injectivity. Fracture permeabilities were varied by changing apertures according to the cubic law that describes the fracture fluid flow. In doing so, it turned out that further hydraulic parameters such as fracture storativity and fracture porosity may be further discussed and investigated. The present-day, undisturbed 3D temperature distribution was modelled taking into consideration the regional temperature gradient, temperatures measured in the borehole and thermo-physical parameters measured in former local geothermal research projects. To access the development performance and development efficiency of the reservoir, a fictive second borehole was incorporated and transient thermal-hydraulic 3D simulations were carried out. In addition, a 3D geomechanical model was developed, which considers a most probable strike–slip stress regime in the 4.5 km deep reservoir by choosing adequate boundary conditions in line with the governing stress field and geomechanical data compiled in the previous project Dolomitkluft and further geothermal projects targeting the Upper Jurassic formation in the neighborhood. The model was calibrated taking into account pressures derived from formation integrity tests (FIT) and stress-limiting constraints. Fault stability studies as dilation and slip tendency analyses were conducted, considering fully coupled thermoporoelastic mechanisms that include the effects of both fluid pressure and temperature changes of the pore-fracture system on the mechanical behavior of the reservoir during different injection and production profiles.

In this contribution we will discuss current challenges in the modelling of the non-linear physics describing relevant thermal, hydraulic and mechanical processes as occurring during reservoir operations. We will detail how these problems are mathematically formulated, based on our, still non exahustive understanding of their couplings and describe how to efficiently implement the resulting physics into open source numerical solutions. In a second part, we will present an in-house hybrid scalable open source software (GOLEM) that can tackle the aformentioned dynamics, with a short digression on its portability to state of the art HPC architectures. This discussion will be complemented with several field study cases, which will help to showcase how such modelling techniques can help in delineating most efficient reservoir protocols both in terms of the long-term sustainability and hazard components. We will discuss how exploitation scenarios can be optimized by feeding field data onto such multiphysics models (providing on the fly solutions) as well as how novel theoretical fracture mechanics concepts can assist to estimate the induced seismic hazard from the operations. We will end this contribution outlining directions where ongoing and future efforts should be focused in the final attempt to arrive at a comprehensive reservoir numerical framework to be used in reservoir studies.