Periodic Reporting for period 2 - HYDRA (Hydraulics modelling for drilling automation)
Período documentado: 2018-03-01 hasta 2020-02-29
Scope of the HYDRA project:
Societal uses for the drilling of deep wells are abundant and without exception have enormous impact on global economies; examples include the exploration and harvesting
of minerals, geothermal energy, oil, gas, and the geological storage of carbon dioxide. The future sustainability of the harvesting of these resources requires the exploitation of
difficult-to-access, unconventional reserves. The drilling of deep wells for these purposes is characterized by high complexity, high uncertainty, high risk and high cost. In
particular, two threats are apparent for the economically feasible and environmentally safe harvesting of such energy and mineral reserves:
1. Productivity: Cost-effective drilling is key to increase the recovery of existing resources. A world-wide trend shows that well construction costs have increased,
while drilling efficiency has reduced. This is a general trend that applies throughout the drilling industry world-wide. Herein, an important factor is the lack of automation
of the drilling operations. Another dominating contributor to reduced productivity of drilling operations is the occurrence of downhole incidents due to errors in downhole
pressure management and the resulting lost time of correcting these errors.
2. Pressure control: Maintaining the downhole pressure within the pressure limits of the well, i.e. between the pore and fracture pressure, is critical for the safety of drilling
operations. If the downhole pressure exceeds the strength of the formation (i.e. the fracture pressure), it will fracture the wellbore, causing a loss of drilling fluid to the
formation, possibly damaging the reservoir. On the opposite side, if the downhole pressure reduces below the formation pore pressure, it will cause an unwanted influx
of formation fluid into the well. This is referred to as a well control incident, which in the worst case can escalate to a blow-out of hydrocarbons on the rig (ref. Deepwater
Horizon as an extreme case).
Development of technology for improved automatic pressure control, such as Managed Pressure Drilling (MPD), is therefore one of the most effective measures to improve
sustainability of harvesting resources. A fundamental prerequisite for automatic pressure control in case of well control incidents is a fit-for-purpose hydraulic model: it needs to
capture the relevant properties of the complex multi-phase flow dynamics of the downhole drilling process, while being fit for model-based control design. To this date, no hydraulic
models exist that are both 1) accurate enough and 2) simple enough to be employed in the context of real-time estimation and control in case of gas influx.
HYDRA scientific objectives:
1. Develop high-fidelity multi-phase hydraulic models specifically for drilling operations that accurately reproduce the dynamics of gas-liquid flow in the well,
enable accurate prediction of down-hole, pore and fracture pressures, and enable the transition between one-phase and two-phase flow.
2. Develop model reduction techniques for the automatic construction of fit-for purpose models that enable numerically efficient simulations of realistic drilling
scenarios, and are suitable for online, real-time pressure estimation and control.
3. Develop model reduction techniques for the construction of fit-for-purpose hydraulic models that enable effective handling of distributed nonlinearities and
delays (due to wave propagation), guarantee the preservation of key system properties (such as stability, multiple time- and length-scales and input-output
behavior), and preserve the dependency on key physical parameters in the eeduced-order model.
HYDRA training objectives:
4. The main objective of HYDRA is to train 3 ESRs to PhD level by means of collaborative research projects. In the scope of the training program, HYDRA
targets both technological and scientific priorities, such as the development of multi-phase hydraulic models and model reduction techniques for drilling
processes, their implementation in software and the testing and validation on industrial benchmark problems.
In summary, the HYDRA EID Network offers a unique research environment where leading academics will integrate ESRs into their research teams for the training period,
providing a top-notch structured training program on both generic professional skills and domain-specific research skills in multi-phase hydraulic modelling, model validation
procedures, model reduction techniques, numerical methods, systems and control, software tool development and testing for pressure management in high-tech drilling
applications.
Societal uses for the drilling of deep wells are abundant and without exception have enormous impact on global economies; examples include the exploration and harvesting
of minerals, geothermal energy, oil, gas, and the geological storage of carbon dioxide. The future sustainability of the harvesting of these resources requires the exploitation of
difficult-to-access, unconventional reserves. The drilling of deep wells for these purposes is characterized by high complexity, high uncertainty, high risk and high cost. In
particular, two threats are apparent for the economically feasible and environmentally safe harvesting of such energy and mineral reserves:
1. Productivity: Cost-effective drilling is key to increase the recovery of existing resources. A world-wide trend shows that well construction costs have increased,
while drilling efficiency has reduced. This is a general trend that applies throughout the drilling industry world-wide. Herein, an important factor is the lack of automation
of the drilling operations. Another dominating contributor to reduced productivity of drilling operations is the occurrence of downhole incidents due to errors in downhole
pressure management and the resulting lost time of correcting these errors.
2. Pressure control: Maintaining the downhole pressure within the pressure limits of the well, i.e. between the pore and fracture pressure, is critical for the safety of drilling
operations. If the downhole pressure exceeds the strength of the formation (i.e. the fracture pressure), it will fracture the wellbore, causing a loss of drilling fluid to the
formation, possibly damaging the reservoir. On the opposite side, if the downhole pressure reduces below the formation pore pressure, it will cause an unwanted influx
of formation fluid into the well. This is referred to as a well control incident, which in the worst case can escalate to a blow-out of hydrocarbons on the rig (ref. Deepwater
Horizon as an extreme case).
Development of technology for improved automatic pressure control, such as Managed Pressure Drilling (MPD), is therefore one of the most effective measures to improve
sustainability of harvesting resources. A fundamental prerequisite for automatic pressure control in case of well control incidents is a fit-for-purpose hydraulic model: it needs to
capture the relevant properties of the complex multi-phase flow dynamics of the downhole drilling process, while being fit for model-based control design. To this date, no hydraulic
models exist that are both 1) accurate enough and 2) simple enough to be employed in the context of real-time estimation and control in case of gas influx.
HYDRA scientific objectives:
1. Develop high-fidelity multi-phase hydraulic models specifically for drilling operations that accurately reproduce the dynamics of gas-liquid flow in the well,
enable accurate prediction of down-hole, pore and fracture pressures, and enable the transition between one-phase and two-phase flow.
2. Develop model reduction techniques for the automatic construction of fit-for purpose models that enable numerically efficient simulations of realistic drilling
scenarios, and are suitable for online, real-time pressure estimation and control.
3. Develop model reduction techniques for the construction of fit-for-purpose hydraulic models that enable effective handling of distributed nonlinearities and
delays (due to wave propagation), guarantee the preservation of key system properties (such as stability, multiple time- and length-scales and input-output
behavior), and preserve the dependency on key physical parameters in the eeduced-order model.
HYDRA training objectives:
4. The main objective of HYDRA is to train 3 ESRs to PhD level by means of collaborative research projects. In the scope of the training program, HYDRA
targets both technological and scientific priorities, such as the development of multi-phase hydraulic models and model reduction techniques for drilling
processes, their implementation in software and the testing and validation on industrial benchmark problems.
In summary, the HYDRA EID Network offers a unique research environment where leading academics will integrate ESRs into their research teams for the training period,
providing a top-notch structured training program on both generic professional skills and domain-specific research skills in multi-phase hydraulic modelling, model validation
procedures, model reduction techniques, numerical methods, systems and control, software tool development and testing for pressure management in high-tech drilling
applications.
Summary of the achievements related to the HYDRA objectives
• HYDRA Objective 1:
HYDRA has developed both single-phase flow and multi-phase flow models for Managed Pressure Drilling (MPD) and has developed a Matlab-based software code for the numerical
simulation of such models. These models and numerical tools have been extended beyond the state-of-the-art by including variable geometry aspects, nonlinear boundary conditions,
coupled reservoir modelling, well-balanced schemes, etc.
• HYDRA Objective2-3:
The ESRs have developed methods for the model reduction of MPD models for the purpose of simulation, observer and controller design.
The type of reduction approaches are:
- Physics-based reduced complexity hydraulics models for both the well and reservoir have been developed.
- Balancing-based reduction approachs for Lure-type systems have been successfully applied to discretized MPD models.
- Data-based moment matching approaches for MPD models have been developed to construct simplified delay system models for MPD systems.
- Balancing-based reduction approaches for delay systems have been developed.
- Reduced Basis techniques for nonlinear systems have been devleoped, applicable to MPD systems.
Furthermore, the results achieved in the context of Objectives 1-3, have been validated in well-defined industrial benchmark scenarios.
THe software developed in the scope of Objectives 1-3 has been transferred to Kelda Drilling Controls and has been embedded within Kelda's existing software for the modelling and simulation of MPD systems.
• HYDRA Objective 4:
Within HYDRA, the ESRs have received an extensive training program on Managed Pressure Drilling (MPD), modelling for MPD systems, numerical tooling
for simulation of such models and on model reduction, through secondments at Kelda, advanced training courses, workshops, coaching meetings.
• HYDRA Objective 1:
HYDRA has developed both single-phase flow and multi-phase flow models for Managed Pressure Drilling (MPD) and has developed a Matlab-based software code for the numerical
simulation of such models. These models and numerical tools have been extended beyond the state-of-the-art by including variable geometry aspects, nonlinear boundary conditions,
coupled reservoir modelling, well-balanced schemes, etc.
• HYDRA Objective2-3:
The ESRs have developed methods for the model reduction of MPD models for the purpose of simulation, observer and controller design.
The type of reduction approaches are:
- Physics-based reduced complexity hydraulics models for both the well and reservoir have been developed.
- Balancing-based reduction approachs for Lure-type systems have been successfully applied to discretized MPD models.
- Data-based moment matching approaches for MPD models have been developed to construct simplified delay system models for MPD systems.
- Balancing-based reduction approaches for delay systems have been developed.
- Reduced Basis techniques for nonlinear systems have been devleoped, applicable to MPD systems.
Furthermore, the results achieved in the context of Objectives 1-3, have been validated in well-defined industrial benchmark scenarios.
THe software developed in the scope of Objectives 1-3 has been transferred to Kelda Drilling Controls and has been embedded within Kelda's existing software for the modelling and simulation of MPD systems.
• HYDRA Objective 4:
Within HYDRA, the ESRs have received an extensive training program on Managed Pressure Drilling (MPD), modelling for MPD systems, numerical tooling
for simulation of such models and on model reduction, through secondments at Kelda, advanced training courses, workshops, coaching meetings.
Key achievements of HYDRA's research beyond the state-of-the-art are:
1. Multi-phase flow (drift-flux) models for Managed pressure drilling including nonlinear boundary conditions (regarding choke charecteristics, bit characteristics, etc.), variable well-bore geometry and coupled wellbore-reservoir models
2. Improved numerical schemes for drift-flux: well-balanced schemes and schemes that can cope with variable-geometry problems.
3. Simulation software for the drift-flux model for managed pressure drilling under point 1, using the schemes under point 2.
4. Model reduction techniques applicable to models for managed pressure drilling, based on a) Reduced-basis methods, 2) Data-based moment matching techniques and 3) Balancing methods.
5. Definition of real-life (benchmark) drilling scenarios with and without gas-influx and validation of the developed models against real-life field data in these scenarios.
1. Multi-phase flow (drift-flux) models for Managed pressure drilling including nonlinear boundary conditions (regarding choke charecteristics, bit characteristics, etc.), variable well-bore geometry and coupled wellbore-reservoir models
2. Improved numerical schemes for drift-flux: well-balanced schemes and schemes that can cope with variable-geometry problems.
3. Simulation software for the drift-flux model for managed pressure drilling under point 1, using the schemes under point 2.
4. Model reduction techniques applicable to models for managed pressure drilling, based on a) Reduced-basis methods, 2) Data-based moment matching techniques and 3) Balancing methods.
5. Definition of real-life (benchmark) drilling scenarios with and without gas-influx and validation of the developed models against real-life field data in these scenarios.