The Costs Tab (figure 102) shows the turbine-specific costs for each turbine of this type. The units are arbitrary. This information is for use in any cost optimisation modules. Per turbine costs are split into two in order to allow a portion of that cost to be subject to a cost multiplier. These two costs together represent the total incremental cost of adding a turbine. It is up to the user as to whether the total incremental cost should be split or not.
When a bathymetry layer has a non-zero value at a turbine location, that turbine will be assigned a foundation cost, based on the water depth at its location, from the table shown in figure 102. The range of water depth for different foundation types should not overlap. The foundation types, ranges and costs show here are given as default examples and of course can be added to, changed or deleted. When a bathymetry layer is not present or has a value of zero, the single value above is used.
Some sources for bathymetry data are
•GEBCO - this is a global dataset at relatively low resolution but it is global and appears to be actively maintained. We recommend downloading the ESRI ASCII grids.
•NOAA - has good high resolution coverage of the US coast although many of the surveys were carried out several decades ago. Openwind recognises the a93 format and will automatically resample and surface it.
•EMODNET - has coverage of all of Europe up into the Arctic including the North coast of Norway and Finland, down past the Mediterranean and along the North African coast, also including the Black Sea and the Red Sea. Again, downloading the ESRI ASCII grids (*.ASC) are likely the path of least resistance when it comes to loading into Openwind.
“Periodic Costs Per Turbine” represents the ongoing maintenance and component replacement costs over the lifetime of the project. Longevity is measured in years; users may add and delete new items by using the "+" and "-" buttons, respectively.
Figure 103 shows a simple interface for defining component replacement times and/or periodic costs based on various measures of site suitability. This can be accessed by double-clicking on a periodic cost. Due to the limitations of this interface, it is likely that single real cost functions would be broken down into several items as in figure 102.
Period costs can be defined with a fixed period and fixed cost. However, these are equivalent to changing the total turbine cost in terms of their influence on the layout. They change the balance between turbine and balance of plant CAPEX and so change the overall importance placed on minimising access road and cable length. This doesn't mean that thee types of periodic costs should not be included but they are not as interesting as those described below.
Varying either the period of the cost with some measure of suitability is a way to bring suitability into the layout design while avoiding Boolean step changes. In order to simplify and de-clutter the interface, the options have been reduced in the latest version to leave only the more meaningful options.
Generally it doesn't make much sense to vary both the period and the cost so we will focus on varying one at a time here.
Variable Cost - This could be used for annual O&M costs in the form of unscheduled maintenance. Annual costs should obviously have a fixed period of 1 year and then we can choose the most appropriate Pseudo-Equivalent Loads (PEL) ratio; either the overall PEL ratio or perhaps a measure of the PEL ratio based upon a specific Wohler exponent. In the latter case, one could add multiple periodic costs with different Wohler exponents and multipliers to represent different types of unscheduled maintenance.
Figure 104: Periodic costs with variable period
Variable Period - The idea here is that we are modelling a component failure. Ideally, if the layout is suitable and the turbine design lifespan is equal to or greater than the project lifespan, then these kinds of periodic costs should never happen. A layout which has 1.5RD separation and horribly waked turbines will result in significantly overshooting the reference turbulence curves and the resulting PEL ratios greater than 1. In the example setup shown in figure 104 above, the design life of the turbine is assumed to be 20 years and a PEL ratio greater than 1 will result in a component failure before 20 years and if the ratio is significantly more than 1 then the lifespan of this component will be significantly less than 20 years. Within the financial model, say we have a discount rate of 10%, a cost that occurs in year 18 will be 10% greater than the same cost showing up in year 19. In this way, with sufficiently large component costs, we can transform suitability from a Boolean or arbitrarily weighted objective factor, seamlessly into the same units used by the rest of the objective function which is to say units of currency per megawatt-hour.
If a loads look-up table is loaded and components have been created, these components will show up in the drop down list after the PEL options. Components are assumed to have the same design life as the turbine itself. The simplest way to setup a component failure cost is to do the following.
Figure 105: Periodic costs using loads component
In figure 105 the PEL ratio has been replaced with an estimated life-span for the low speed shaft. In this case the lifespan is 20 years which is the same as the turbine but we could easily change it to 30 years by changing the multiplier from 1 to 1.5.
The Days Lost parameter is triggered whenever the period is realised as a drop in availability for the year in which the periodic cost is applied and is intended to reflect the amount of energy production time that would be lost while repairing the turbine. In order to implement this, the annual production is modified for each year to take account of a decrease in energy for the specific turbine or turbines which are losing days of production. This is implemented as a simple drop in availability for the turbine undergoing maintenance but other complex compensatory effects (such as a decrease in wake effects or fatigue loading at neighbouring turbines whilst a turbine is undergoing maintenance) are not considered. The days lost input can be ignored by setting it to zero.