OZ-ENERGY-ANALYSIS.ORG   -   open science for the new millennium

THE STORIES   |   DATA   |   ANALYSIS   |   MODELS   |   LITERATURE   |   DISCUSSIONS

Status: testbed work from November 2010

Simulate Wind Farms on Eyre Peninsula from BoM wind speed data

ABSTRACT and INTRODUCTION

Here is our first pass at simulating Wind Farm output from Bureau of Meteorology (BoM) wind speed data. The Eyre Peninsula is of particular interest, as per the SA Green Grid Proposal. What we develop here will be a starting point for the larger Broome to Cooktown Challenge, where we will explore the potential output from wind farms on a national scale.

The Eyre Peninsula is a useful test bed as there currently exist two wind farms (Mt Millar and Cathedral Rocks) in this region, and the known output from these will be useful in judging the results of the simulation.

DATA and METHODS

The wind speed data used here is described on its own page: Eyre Peninsula Wind Speed Data.

The method (the Wind Power Curve model) for transforming BoM wind speed data into putative wind farm output is developed on the Reconcile Wind Farm output with BoM wind data page; and, in particular, is formalised in these Matlab scripts (which can be translated into other languages / forms, taking care to handle missing data).

The output data for the existing Mt Millar and Cathedral Rocks wind farms can be found on the Wind Farm Generation Data page.

ANALYSIS and RESULTS

With the data and methods as above, it is straightforward to transform the wind speed data into simulated wind farm output, assuming a capacity factor of 35%. The following pseudo code does this job:

load '/OzEA/data/wind/EP_wind/EPwind_R1_60min_2009.csv'

DPpD   = 24;     % Data Points per day 
DYst   =  0;     % Day of Year start
CF_req =  0.35;  % Capacity Factor for simulated wind farms

for i=1:10,
    wind     = EPwind_R1_60min_2009(:,i);
    wind_cor = OzEA_wind_speed_corrections( wind,  DPpD, DYst );
    sim(:,i) = OzEA_simulate_wind_farm_output( wind_cor, CF_req );
end

The resultant simulated wind farm data ('sim'), output and formatted up, is here given as: OzEA_R1_WFsim_EP_60min_2009.csv (640 KB).

There are numerous sanity checks and exploratory analyses I want to do now, but... need to cut to the chase.

[Friday eve] Ok, been kicking this around, trying to work out what the point is here... and the point has been to kick things around in preparation for doing much the same analysis on the Broome to Cooktown set. So, will now step through a few things. First up, here is a plot simply showing the data we have simulated (albeit boiled down to daily averages):

We see, as now expected, and at this courser level, that the power output from the wind farms goes up and down as the weather passes across. A detailed view of what this looks like (albeit with somewhat different data) has long been available in the data viewer - and there seems little point in providing such a view here. We need to get more abstract, to boil this data down into a view that provides quantitative hooks into a larger synthesis... [as now given in the demand remainder plot of comment 3]

REMARKS

[preliminary]   The detail of how this analysis is performed was built elsewhere (see Data and Methods), and the resultant simulation data traces are, on the face of it, unsurprising. Development of this page can proceed in the comments, until such time as more specific remarks here are warranted.

3 OzEA_ASWFEP0003

francis Subject: renewables and the demand profile Date: 2010-11-30 (at 01:47:25)

Simulated Eyre Peninsula wind farm data is used in a simple analysis of the SA power supply. Of course, SA is in reality connected into the NEM; however, it is somewhat convenient to remove this aspect here, in part because 2 GW nameplate of Eyre Peninsula wind power is almost half the SA market in average terms, and it is ultimately one half of the NEM market that we will model as powered by renewables.

As in the following plot, we use cumulative analysis more and more regularly, so it is worth taking a moment here to digest. The x-axis shows the level of supply (or demand), and the y-axis shows the percentage of the time that a given level of supply is exceeded.



The blue overall SA demand shows clearly where the old-fashioned idea of "baseload", "intermediate load", and "peaking supply" come from. The first 1 GW is a constant requirement; the next GW is 'intermediate', cutting in during the day and then cutting out at night; the tail above 2 GW (up to 3.4) shows the extremes, especially as occurring on particularly hot days (i.e. more occasionally than what I have characterised as the daily cycle of intermediate loads).

If the wind power were a perfectly constant source, then, at 35% Capacity Factor, it would provide a constant 700 MW. We can simply subtract this from the overall demand (red curve) and the result is trivially a 700 MW shift taking up most of the 'baseload' and leaving the 'intermediate' and 'peaking' loads as they were. Of course, the wind power comes and goes, sometimes for days at a time, and so we use the simulation data to examine how this changes the demand profile.

Here I used three BoM stations equally (18116, 18191 and 18083) to represent an Eyre Peninsula with 2 GW nameplate of wind, and again subtracted this wind power from the total SA demand (hour by hour) before doing the cumulative calculation and plotting the profile of the 'demand remainder' (green curve). In a very real way the old-school "base supply" has been taken out by the wind power, and the 'intermediate' and 'peaking' phases have changed shapes. We will focus more on this way of looking at the inclusion of renewables in the coming weeks and months.

Note that the preliminary model from mid year has a very similar plot of the demand reminder profile, and goes further by examining how storage can change the profile shape.

[Show Full Lists]

 

  TOP   |   REFRESH PAGE   |   ANALYSIS

fc - November 2010