The rapid growth in world population and increasing demands have led to the lack of fresh water, taking a toll on several of earths reserves, mainly the fresh water supply reserves. Water stress deters economic growth, leads to conflicts and has a direct impact on the health of humans. Studies show the trends in the
increase of water consumption per capita due to increase in higher standards of living over the last years, resulting in the decrease of usable high purity water. 97% of the world’s water supply is locked in the salted, often unusable oceans. In recent times, water stressed countries are using the saline water from the oceans and desalinating it to produce fresh water for domestic, industrial or agricultural use. The state-of-the art method for desalinating saltwater is by Reverse Osmosis (RO). The biggest drawback of this technology is its high energy consumption mostly provided by conventional sources like fossil fuels. Therefore, for a sustainable future, a renewable energy source must be integrated to power the RO system.
Delft Offshore Turbine (DOT) is currently developing and testing a new hydraulic drive train solution for fluid power transmission in offshore wind turbines using seawater as the medium. A hydraulic positive displacement pump is driven by the DOT wind turbine creating a flow of sea water under high pressure. This high-pressure flow can be either directed to a RO unit to desalinate seawater or converted into electricity using a spear valve and pelton turbine generator. A major challenge while using wind as an energy source is its intermittency. Reverse osmosis plants are designed to operate at a fixed flow and pressure while due to the uncontrolled, varying nature of the wind, the pump output experiences fluctuations in both pressures and flows.
The aim of this thesis is to analyze the steady-state behavior of the integrated system for wind speeds up to the turbine rated wind speed under different operating conditions. This is achieved by simulating the behavior of the DOT500 hydraulic drive train wind turbine coupled to a RO system with pressure exchanger energy recovery device and a nozzle with a pelton turbine generator. The main objective of this research is to build a numerical model in Python using algorithms to solve the system of steady-state equations and optimize the integrated system for maximum water and maximum electricity production at specified locations.
The numerical model is simulated for all combinations of wind speeds and nozzle positions and the behavior of the high-pressure pump, RO system, pelton turbine and various system parameters are shown. A sensitivity study is performed for important desalination parameters and their effect on system performance is analyzed.
This research has yielded three main conclusions / deliverables:
1. Behavior of the wind powered integrated system to produce freshwater and electricity at different operating conditions.
2. Steady state control of DOT500 turbine and behavior of the pelton turbine to produce either maximum electricity or maximum water.
3. Sensitivity of important desalination system parameters on overall system behavior for future optimization of RO membranes.

The modeling method for the performance of positive displacement pumps of hydraulic pumps is based on collecting operational data of the pump for its operational envelope and using equations with empirical fitting parameters and fit these equations to the data. This way values for the fitting parameters are obtained, with which the equations can be used as model for the operating performance of this specific pump. By implementing this modelling method no information about the contribution of each component present in the pump to the performance losses is obtained. A model which analyzes each coponent and estimates its contribution to the total power losses would be beneficial in the design of pumps. Therefore the answers to the following research questions are sought after in this work:

- What are the components critical to performance within hydraulic pumps?
- To what accuracy can a component-wise, analytical model for pump performance be created?
- How sensitive to certain parameters used in the equations is the model?
- What are the possibilities of implementing this model into pump design?

After the pump system was analyzed and the internal interactions found, the components which effect pump performance the most were identified. These include bearings, rollers, seals, springs, valves and leakage flow which is dependent on gap size, piston movement and viscosity of the hydraulic fluid. For these components analytical equations are presented which are incorporated in the model.

The accuracy of the model is tested by comparing its results to test data of an experimental camring driven radial piston pump. The model predicts leakage flow to an average difference of 3% relative to the data, but outliers up to 49% for high pressures and -35% for low pressures are observed. Input torque is predicted to an average difference of -19% relative to the data. This difference is mainly caused by the inaccuracy of the predictions below 30 bar operating condition, where outliers up to -61% are observed. These results are compared to an empiric modelling method and are found to be more accurate for leakage flow for the entire operational envelope, yet on average less accurate for input torque.
The sensitivity of the model to certain parameters is tested. The parameters investigated are the rolling resistance coefficient of rollers and bearings, bearing lubricant viscosity, hydraulic fluid temperature, piston-cylinder alignment and the corresponding forces resisting piston motion and the assumption of laminar leakage flow.

To show the possibilities of the model in pump design the rollers and appurtenant bearings present in the pump are replaced with hydrostatic bearings. The design is discussed and a hydrostatic bearing with an orifice restrictor is designed of which implementation will approximate the total efficiency of the pump with rollers. The Reynolds number of the flow necessary is found to be too high for capillary restrictors. To approximate the same total efficiency for each operating condition the average bearing flow is found to be 0.0016 L/s, which corresponds to an average fluid film height of 12 μm. The orifice restrictor diameter which allows this flow is found to equal 0.24 mm.

This work shows the possibilities of such a model, yet there is plenty room for improvement of the accuracy of the predictions made. Recommendations to this end are given in the conclusion and recommendations of the report.

As wind turbines continue to increase in size, so do the loads acting on the drivetrain.
To handle the higher loads the mass of the drivetrain needs to increase. A better understanding
of this mass increase can result in lighter drivetrain design and provide
an insight which drivetrain design has the lowest mass potential for large rotor diameters.
In this thesis the limiting mechanisms, mass and torque density of three
wind turbine drivetrains are investigated for an increasing rotor diameter. The limiting
mechanisms in this thesis provide a theoretical limit to the minimum required
mass of a drivetrain for increased loading. Through an extensive literature study the
limiting mechanisms are found and three scaling models are developed to calculate
the torque density of the drivetrains. The drivetrains that are being considered in this
thesis are the drivetrain with gearbox and high-speed generator, the direct drivetrain
and the hydraulic drivetrain. The following research questions are answered in this
thesis:
• What are the limiting mechanisms of the three investigated drivetrain types?
• What are the achievable torque densities of the three investigated drivetrain
configurations for increasing rotor diameters?
Two limiting mechanisms are found for the drivetrain with gearbox and high-speed
generator: Tooth flank stress (1.72 GPa) and Root bending stress (0.24 GPa). The
Tooth flank stress is found to be governed by the Hertzian contact strength while the
root bending stress is governed by the flexural strength. Two limiting mechanisms
for the RFPMSG are found for the airgap flux density (1 T) and the current loading
(30 - 200 kA/m) in the generator windings. The airgap flux density is limited by the
saturation of the stator teeth material in the generator whereas the current loading
is limited by the heat dissipation of the windings and stator laminations. For the
hydraulic pump, the limiting mechanism is found at the interaction of the camring
and the cam roller. The maximum stress that is allowed at this point is determined
by the Hertzian contact strength of the material and is found to be (1.72 GPa). Using
the limiting mechanisms of the three drivetrain configurations, scaling models for
the mass and torque density of the three drivetrains have been developed.