NWTC met towers

On the NWTC there are several sources of meteorological data. NREL operates two towers;

• The M2 tower. This has sensors up to 80 m a.g.l. and is the reference data set for the site, having been operational from the mid 1990's on. It's about 250 m WSW of the HRDL at the moment, an 1 KM west of the large turbine test stands on row 4. We have continuous data from this tower, and it's publicly available online at http://www.nrel.gov/midc/nwtc_m2/
• A tower in front of the CART turbine, which is the two-bladed turbine on row 4. When the CART is not operational, we have data coming off a 58m tower, including a sonic at 36 m. This tower also has wind speed, direction an temperature at 3, 15, 36 and 58 m. We're working out what data is available during the TWICS IOPs.

We're in the process of commissioning a new, 135 met tower at site 4.4, directly upwind of the turbine. That tower was instrumented earlier this year, and the termination to our data acquisition systems was completed in April. Since then, we have been working to check the data coming off the tower. For those who aren't familiar with this, the data collection process goes something like this;

1. The physical property of interest is converted into a signal of some sort by an instrument. So, wind speeds are measured by cup anemometers, which generate a pulse every time they rotate. The speed of rotation is then a function of the wind speed. Wind directions are measured by vanes, where the rotation of the vane into the wind changes a resistance and so we get a voltage signal which is proportional to wind direction. And so on, for temperatures, pressure and whatever else is of interest.
2. These signals are transmitted down the tower to our computers, where we use a variety of hardware to convert the signals into something that a computer can read.
3. Then we scale the signals that the computer reads, back into so-called 'engineering units'. This is the opposite of what is done by the instruments; we convert voltages back to wind directions, frequencies to speed, and so on.
4. Then we save the data on the local computer. To keep file sizes down, we use a binary format. We have more than 70 data channels that we record at 20 Hz, so this is an essential step.
5. Next, data moves to the server. We now have two copies of our data in seperate locations.
6. Then we load the data and reduce the 20 Hz data to 10 minute data so that we can get a quick look at what's happening, and save this data for reference.
7. In future, we can revisit the high-frequency data for interesting events.

So, 'commissioning' means that we check every step of the above process. Are our voltages and frequencies right? Are we picking up the right piece of data and associating it with the right instrument? Do our files look right? And so on, ad nauseum.

Let me give you a flavour of where we are now, and the kind of things we are dealing with.

1. The wind vane orientation is being checked and adjusted. This means that any data from earlier will have to be adjusted slightly, but by an amount that we can quantify.
2. Accelerometers that show how much the sonic anemometer booms are vibrating are showing a drifting offset from the values they should show. I'm not going into the technical reasons for this, but it's annoying.
3. We're checking the process of converting data to binary and back again.

In terms of instrumentation and data, we have the following

1. Sonic anemometer data should be OK by now, but we have questions about how large the boom vibrations are.
2. Wind speeds look OK, but wind directions are not calibrated at all heights (we've done to 26 m a.g.l up to now).
3. Ground pressure is fluctuating quite severely, which may or may not be realistic.
4. Temperatures look reasonable.

I'll provide updates as we go along, but it's worth keeping in mind that this kind of thing can last a while. My gut feeling is that we're close, although there are always things that pop up when doing this kind of work.