PETROMAKS2-Stort program petroleum

To reduce the environmental impact from the Norwegian petroleum operations, produced waters are often re-injected in dedicated shallow formations, or re-used for Enhanced Oil Recovery purposes. Such injection procedures pose the risk of leakage, or even cap-rock failure, potentially following misinterpreted pressure management. Microseismic events are a direct and early indicator of geomechanic deformation of the reservoir due to stress and pressure changes. New fractures generated by these (micro-) earthquakes offer new leakage pathways for the injected fluids. Early identification of potential leakages can thus help to develop mitigation strategies. In particular, these early indicators are essential in the fragile environment of the Arctic reservoirs of the Barents Sea. The StressLess project aimed at researching into how to combine the two key technologies, microseismic and geomechanic modelling, into an earthquake early warning system. Two classes of microseismic events can be defined as critical for early warning systems: Class I events are (i) fluid-driven; (ii) within the reservoir; (iii) mostly tensile opening fractures; and (iv) show typically b-values larger than 1. On the other hand, Class II events have fundamentally different characteristics: (i) stress triggered; (ii) occur at some distance from the reservoir; (iii) mainly pure-shear; and (iv) have typically b-values around 1. Class II events typically contain the largest magnitude events observed. While Class I events enhance the storage capacity of reservoirs, Class II events rather pose an environmental risk. The first objective of the StressLess Project was to develop data-driven methods to discriminate between Class I and Class II microseismic events. Automated, machine-learning methods provided objective, data-driven criteria based on various waveform characteristics and seismological parameters. The second objective was to establish geomechanical models, able to distinguish between the two event classes. As such, the models need to include direct physical coupling between pore pressure and stress evolution in the rock matrix. Furthermore, the models must allow fractures to occur without including a priori failure planes. A volume conservative geomechanical numerical framework was developed. The numerical implementation was validated against an analytical model for an impermeable and homogeneous rock, where the critical fluid pressure is constant. Implementations for both permeable and low-permeable rock were developed, based on the concept of "invasion percolation". We found that "high" permeabilities of the damaged rock give the same b value approximately 0.8, but "moderate" permeabilities give higher b values. Another difference is that "high" permeabilities produce a percolation-like fracture network, while "moderate" permeabilities result in damage zones that expand circularly away from the injection point. Moving to recorded data from offshore Norway, we developed a method to retrieve moment tensors from Ocean Bottom Cable records. This is an additional important parameter to assess the risk of a specific event, as well as a group of events. Furthermore, this rupture mechanism adds valuable constraints on the geomechanical models.

- NORSAR's work on the offshore data has directly contributed to the proposal development of the SNSnet project. - Experience gained in StressLess regarding moment tensor inversion will give valuable constraints on future event interpretation. The work already confirmed a casing failure recorded as event. Quick identification of the event type has direct impact on decision making for both production and injection, and thus risk mitigation

To reduce the environmental impact from the Norwegian petroleum operations, produced waters are often re-injected in dedicated shallow aquifers or re-used for EOR purposes. Such injection processes and general pressure management pose risks of leakage or even cap-rock failure. Misinterpreted pressure management may result in rapid deformation, often manifested as micro-earthquakes or even larger magnitude earthquakes. The three main knowledge needs targeted in this proposal are: 1)we need to know how to characterize microseismic events such that we can quickly discriminate events to belong to class-I (pressure induced events) or class-II (stress triggered events) type of events 2)we need to understand through geomechanical modelling under which circumstances we can expect class-I and class-II type of events 3)we need to increase the competence in integrating real-time microseismic monitoring with geomechanical modelling such that there is a feedback loop between monitored deformation and stress field modelling. Solving these knowledge gaps, we will get closer towards workflows for early warning systems towards leakage detection and pressure management in particular during re-injection of produced water. The proposed project consists of three scientific work packages that will be based on the analysis of high-resolution industry datasets. The main dataset will be provided by Statoil and will consist of passive seismic data from an ocean-bottom cable system offshore Norway (likely Snorre or Grane fields). From the same field we will receive all relevant geomechanical model information. The secondary dataset will also be provided by Statoil and will consist of passive seismic data from onshore hydraulic fracture stimulation. This dataset will serve as a controlled test experiment for both microseismic analysis and geomechanical stress modelling, before transferring the gained knowledge to the offshore Norway case, where noise conditions pose an additional challenge.