Artificial vision systems for bird impact

One of the problems of installing a wind farm in an area with a dense bird population is the possibility of impacts between the rotating blades and the animals.

A standard solution used in the industry is the use of a (very expensive) bird radar, a quasi-military technology that can spot very small flying objects and stop the turbines (hopefully) in time.

Working at the development of a wind farm in Jordan we have discovered the existence of an alternative: a Spanish company (Liquen) is developing an “artificial vision” system to be installed on the WTGs.

Basically the system is composed of 2 sets set of high definitions cameras installed on the WTGs or on the MET mast that are filming the space around the turbines, plus speakers for dissuasion sounds.

The software can analyze the images in real time and recognize if a bird is flying toward the machine. If this is the case, several countermeasures are possible:

  • Warning and dissuasion, using annoying signals. Clearly this signals are optimized to the type of birds in the area
  • Stop control, whit short duration (<2 minutes) stop of the affected turbine and automatic restart to minimize production loss

According to the company, real time detection is really fast (less than 1 second) and it is possible including in adverse weather conditions (fog, snow, rain) whit a very low power consumption.

The main problem that I see is that the system is effective only when there are at least 200 Lux (that is, from sunrise to sunset). So there is a clear problem with nocturnal birds – I wonder if in the future it will be possible to integrate the system with some kind of night vision, like the military infrared technology.

Another weak point is the low detectability of very small bird (12 centimeters or less): they can be easily spotted only if they came in a large group.

This technology looks promising and it’s not so “embryonic”: it has already been installed in a Vestas wind farm in Greece and several other wind farms around Europe (Spain, Norway and Greece).

If you are interested you can find more information in the DTBird Brochure or in the results of a study in a real wind farm in Norway.

Relevant parameters in wind farm production

After the connection of a Wind Farm to the grid several parameters are used to analyze the smooth operation of the installation.

The more relevant are:

Capacity Factor

CF=\left( \frac{E(kWh)}{P(kW) T(h)}\right )

Is a parameter used frequently in power producing  plant. A high capacity factor means that the plant is working almost continuously (for instance a nuclear plant), while a low capacity factor may characterize a power plant working only in peak hours (like some hydro plants).

In the case of wind farms, capacity factor depends more on the wind that on the needs of the grid.

To be economically reasonable, a wind farm needs to have a capacity factor of more than 25%. Translated in hours, it would be around 2190 equivalent hours.

This parameter is probably too “global”, as it doesn’t add information about why the wind farm was not producing: was it for a low wind, for a technical problem of the WTG, for a disconnection from the grid?

Or maybe it is due to a scheduled maintenance or to the wind sector management (the automatic planned disconnection of some WTGs in particular wind conditions)?

Technical Availability

TA=\left( \frac{T(Available)}{T(Total)}\right )

This is easily defined: basically is the ratio between the hours the WTG was available for production an the total number of hours in the considered period. If there is a fault in the grid, or if wind condition is above or below the maximum, it doesn’t count as “unavailability”.

Production Availability

PA=\left( \frac{T(Producing)}{T(Total)}\right )

This is parameters start to be interesting from an economical point of view: is the ratio between the total number of hours producing and the hours in the considered period. It will be less than 1 due to grid disconnections, WTG problems and wind outside the operational limits.

Effective Availability

EA=\left( \frac{T(producing)}{T(wind)-T(disconnected)-T(stop)}\right )

This parameter give a very solid information about the quality of the turbine, and the “real” availability: is the ratio between the hours of production and the hours of wind speed between the operational limits, minus the hours disconnected by the grid (for a grid problem or order) minus the justified stops (for visits, preventive maintenance, etc.)