Glasgow Analysis
System Inputs and Parameters
From the three Scenarios that were analysed in this case study the approach for the first two was similar in the way data was input to HOMER.
Vestas V80 Wind Turbine
The first stage was creating a power curve for the Vestas V80 turbine used, created using the Danish wind site [1]. This was input to HOMER in the wind turbine block. The wind resource for the ardrossan site was then synthesised using Scottish wind data provided by Sgurr_Energy and scaled up in HOMER using an average wind value of 7.4 m/s, found using the online BWEA wind speed database [2]. The power output of the wind turbine was then calculated by HOMER using these two parameters in hourly intervals over the course of the year.
Electrolyser
The alkaline electrolyser used in the buffering system was simply modelled with a constant efficiency in HOMER. This efficiency differed from the other efficiencies in HOMER in that it was calculated at the HHV of hydrogen of 142 MJ/kg [3]. A ratio of this and the LHV of 120 MJ/kg was then taken to factor the efficiency value by and calculate a LHV efficiency of approximately 59%. This value was taken to be representative of a large bank of electrolysers that would be required for the buffering systems used in this case study. Other input parameters for this block included a lifetime of 20 years, which was taken from the upper end of the current electrolysers available on the market.
Compressor
There is no compressor programmed into HOMER therefore Microsoft Excel was used to model the compressor. The first law of thermodynamics was used to calculate the unit work requirement for compression and the power required. This approach was based upon a project undertaken by the European Commission and Institute for Energy [4] and is outlined below:
- The first value to be calculated was the maximum hourly hydrogen production rate via electrolysis:
Where:
= Efficiency of the electrolyser plant (%)
= Lower heating value of hydrogen (kJ/kg)
= Density of hydrogen at normal conditions (kg/ Nm3)
a = Conversion kJ to kWh (1/3600 kWh/kJ)
• Following from this the theoretical compressor work required was calculated:
Where:
= Theoretical compression energy required (kJ/kg)
= Adiabatic exponent = 1.41 for diatomic gas Z = Hydrogen compressibility factor at inlet pressure (dimensionless)
= Hydrogen gas constant (kJ/(kgH2.K)) 
= Inlet hydrogen temperature (K) 
= Required hydrogen outlet pressure 
= Hydrogen pressure at compressor inlet
It is assumed that the hydrogen pressure at the compressor inlet, p0, is the same as the
pressure of the hydrogen leaving the electrolyser unit, thus, pressure losses are not taken into
account.
• From the theoretical work, the actual unit energy requirement for compression could be calculated:
Where:
= the efficiency of the compressor
• Finally a rated power value for the compressor was calculated:
Compressed Hydrogen Storage
The hydrogen storage component of the system only required two inputs in the form of a lifetime in years and an initial tank level. The lifetime was chosen to be the duration of the project, twenty-five years, because of the increasing use of carbon fibre in compressed storage that leads to increased lifetimes. An initial tank level of 50% was opted for to reduce the size of tank required. The electrolyser and storage tank were then sized upon the criteria of not reaching zero storage level and ensuring that at the end of the year’s simulation the storage level was approximately what 50%. This then guaranteed that there was enough hydrogen to start the next year with.
Fuel Cell
The alkaline fuel cell used in the buffering system is modelled in HOMER as a DC generator fuelled by hydrogen. The lifetime for this component is input in hours and after a degree of market research it was decided a lifetime of 40000 hours was reasonable. A fuel cell curve was constructed in HOMER as described in Case_Study_1_Analysis. This meant that the fuel cell efficiency varied slightly with output and had a mean efficiency of approximately 65%. The most important parameter of the fuel cell that was utilised for this case study was the fuel cell schedule. This was adapted through the use of a user friendly interface that allowed you to set the fuel cell to be forced on, off or running at an optimised configuration (optimisation process described in Case_Study_1_Optimisation Process section). The schedule was arranged by month and was in hourly intervals. This resulted in an inherent limitation in that only month-to-month variation in the fuel cell output was possible. This was the main limitation that lead to the third scenario being simulated purely using Microsoft Excel.
Scenario 1: Constant Annual Power Output
The first of three system configurations analysed for this case study was modelled around the objective of meeting a constant, flat load all year round. This therefore required sizeable storage and electrolyser components because of the need to effectively meet a constant load all year. The wind farm alone has an annual energy output of 86 GWh, which is subject to a round trip efficiency loss from the buffering system. This round trip efficiency is the term given to the efficiency of the entire system from input to output and is calculated as follows:
Where:
is the efficiency of the H2 storage tank
is the efficiency of the fuel cell
The efficiencies of the aforementioned components were calculated at the LHV of hydrogen of 120 MJ/kg and based on current market efficiencies.
This energy loss permits a maximum annual fuel cell output of 28 GWh. This then limits the load the fuel cell can meet each hour to 3.2 MW. This value was rounded down to 3 MW for the fuel cell rated power, giving an annual energy output of 26.28 GWh. In order to manipulate the system to ensure the load was fully met by the fuel cell a 3 MW load was set and the fuel cell schedule was set so that the fuel cell was forced on at all times. In addition to this no inverters or rectifiers were used in this system because of the way HOMER tries to minimize excess electricity by supplying the load with a combination of fuel cell and wind farm power. With this simplifying assumption that no conversion is needed the load can be fully supplied by the fuel cell. An annual energy distribution the system can be seen in the Sankey diagram below.
The result of the load being at a constant level for the entire year meant that the fuel cell was running for the whole year. This resulted in the fuel cell having to be replaced five times through the duration of the project.
Scenario 2: Four Hour Peak Time Power Output
This system configuration produces flat power output in four hour intervals during the peak sellback period of the day. To achieve this, a fuel cell schedule was set to operate four hours a day during the period that coincided with peak_grid_sale_times . This was 17:00 - 21:00 during the first and last quarter of the year and 11:00 – 15:00 during the rest of the year. With this schedule set the fuel cell was able to achieve a power output of 20 MW for four hours a day for the winter months and a power output of 18MW during the same periods in the middle 6 months. This averages out at a power of 19MW for four hours a day, which gives an annual energy output of 27.7 GWh. The energy flows of this system are depicted in the Sankey diagram below:
The hydrogen storage level of this system follows a slightly different trend from that of the first scenario. This is due to the difference in fuel cell output in the summer and winter seasons. The storage of this system was also required to be more than twice that of the first system. This was because the hydrogen store had to supply, on average, 3500 kg of hydrogen a day to the fuel cell. This meant that during the summer months, when small amounts of hydrogen were being generated by the electrolyser, a seventy tonne drop in stored hydrogen occurred.
Scenario 3: Four Hour Intervals Following Wind
This system configuration consists of a fuel cell output that follows the trend of the wind farm output. This could not be achieved through HOMER due to the fuel cell scheduling function of the program. This was done monthly and could not be altered daily therefore the wind farm output data was taken from HOMER and exported to Microsoft Excel to continue the analysis. For this scenario efficiencies were modelled by constant values, in comparison to the fuel curve HOMER uses to model the fuel cell efficiency. The fuel cell input was then calculated as an average every four hours of the wind output after the efficiency losses brought about by the electrolyser and compressor. To evaluate the fuel cell output a fixed efficiency of 65% was used to multiply the input by. Hydrogen storage over the year was then calculated using an hourly energy balance between compressor output and fuel cell input. An energy balance for this system can be seen in the Sankey diagram below:
The trend in stored hydrogen for this scenario is noticeably different from the first two. This can be mainly attributed to the energy balance being done in Excel rather than HOMER. Nevertheless there is still a similarity in trend, in that the supply drops to a minimum in September, at which point there is a sharp spike in the supply. The trend appears steeper than the other two models, though the magnitudes of the spikes in storage level are much smaller in magnitude than the first two scenarios. This is because the amount of hydrogen store is considerably less, requiring only 1500 kg of storage. The steep gradients in the trend are therefore magnified in comparison to the other two scenarios. This reduction in required storage is because of the constant production and draw of hydrogen from the store. A degree of system control was also implemented in this scenario in the form of an IF statement in Excel. The IF statement simply ensured that when there was no wind supply, both electrolyser and fuel cell were not in operation. This allowed the storage size to be reduced and prevented a continuous downward trend in stored hydrogen brought about by a constant draw from the fuel cell.
References:
[1] www.windpower.org
[2] http://www.bwea.com/noabl/index.html
[3] http://www.woodgas.com/hydrogen_economy.pdf
[4] S. Shaw, S.D. Petevs, Bridging the European Wind Energy Market and a Future Renewable Hydrogen-Inclusive Economy, 2006
[5] http://www.e-sankey.com/en/
