Under-performing Wind Model

I'm trying to blog regularly so here goes even I don't have anything terribly exciting to say.

As part of my Honours year I need to do a research unit. While we won't have an exam, we do have 4 major assessments and two of them are due in the next few weeks - eek. As a result, my focus this week is not so much on climate modelling (which, since you're asking is not going too well thanks - but more on that later) rather on the leaf traits of plants in Jurien Bay. Its a bit of a departure from the work I am doing and I must say its nice to be working on something that I can actually achieve a result from.

As for TAPM and wind modelling, my models are still chronically underestimating wind speeds. This is so very frustrating. To try and figure out where my model is going wrong, I make one small change to the model, run it and then need to wait a day for it to finish before I process the data to see if the change has worked. So far I've had no luck so cross your fingers and toes that tonight's iteration of the model is THE ONE.

x

Wind Data

I managed to get back to Energising South East Asia today for a few hours and was lucky enough to meet some people who may be willing to share some 50m wind data with me. This is great because access to this data will really help to add credibility to my project.

In other news, I re ran my model simulation and imported the data into Matlab. Everything was working perfectly. I generated a wind plot from my model and then one from BoM data and they even looked the same! Then I looked at the scale of the graphs......different, dam it!

I have no idea why my model hasn't worked but I'll be re-checking and re-running everything again tonight to try and find out.

Energising South East Asia

I was lucky enough to spend the afternoon at the Energising South East Asia Conference staffing the Office of Energy booth. It is a long time since I have been to a conference and I had forgotten how motivating it can be to spend time with a bunch of like minded, passionate individuals.

The conference brought together a large number of those involved in renewable energy here in WA and has given me the opportunity to put faces to the voices and names of some of the people I have been in touch with over the last few months in addition to meeting many new contacts.

Although I'm not scheduled to go to the conference tomorrow, I am hopeful that I may be able to drop in again at some stage as work is right next door.

Today's Update

Today started really badly. I has big plans to verify my model only to discover that the weather station data I needed didn't exist. What this means is that I have to do another model run, this time centered around a weather station which is logging hourly wind speeds, and start the process again. The good news is that I can still use all the scripts that I have written so once the model has run, the process of converting, loading and analysing the data should be easy.

It also became apparent today that in order to make my research useful, I am going to to need some met data from wind farm sites. Weather stations only measure wind speeds at 10m and I need to get my hands on some actual wind data at a height of about 50 - 60m. I have started getting in touch with some wind farm operators to see if they have data that they would like to share and I also plan on trying to meet up with some wind farm operators when I am at the Energising South East Asia conference tomorrow.

What I'm Doing Now

Two posts in two days - I hope you're impressed!

Yesterday was all about what I have been up to over the last month. Today is all about what's going on now. The most important thing I'm doing is dedicating more time to this project. Despite warnings from several people, I was woefully under-prepared for the time that is needed for a year long research project. Surprisingly, it isn't something that you can just squeeze in between the rest of your life. Can you imagine?

Currently I am setting up two ten year model runs in a CSIRO developed application called TAPM. TAPM (The Air Pollution Model) is actually a pollution modelling tool but I am using the meteorological aspects of the application as a simple (relatively speaking) wind prediction tool. Hopefully I am going to have these ready to run on the supercomputer by early next week and they should be done a few days after that.

At the same time, I am also researching the accepted statistical methods for verifying my model. Stats isn't really my thing but it is unfortunately necessary. So far I actually understand the stats that I need to use which is a pleasant surprise given how complicated everything else has been!

Oh yeah, and one of these days I'm going to get some results to show you however I am going to need to get Matlab to work first... It could be a while.

Back Again

When I started blogging about my research project it was with the intention that I would be posting regularly and I haven't made it far into the year before failing at that. The truth is that the last month and a bit (since my last post) have been a bit arduous really in terms of my project.

Here's what I have been up to while I have been away:

1. I finished my project proposal. It felt very good to get this done because it has helped me to define in broad terms how I am going to spend my time over the next 36 or so weeks.

2. I set up and ran a weather model run! This first step was surprisingly easier than I thought and I generated a bunch of data very quickly. Little did I know then that creating the data is the easy part. Doing anything useful with the data is where it gets hard, really hard.

3. I have spent s**tloads of time trying to figure out how to convert data files, write loader scripts and generally figure out how the bloody hell Octave and then Matlab actually work. After many, many, many hours of beating my head against a wall I am beginning to make progress in this area. I haven't yet done anything useful with my data but I'm getting there.

In summary, I think I'm turning the corner from having no clue to having a small clue. Time will tell..