Wind Turbine Gossau

Client: envergate ag, CH-9326 Horn TG
Report: 100713, July 13, 2010
Publication: 2011 SGEB-DACH Hannover (in German)

Problems arose when operating a small (10 kW) wind turbine being equipped with a three-bladed rotor running around a vertically oriented axis (Fig. 1). The mast vibrations ran out of control as soon as the generator was activated upon the frequency of rotation having reached the necessary level. This frequency increases with increasing wind speed. The problem was tackled using a multiple phase approach. Firstly, a combined analytic/experimental modal system analysis was undertaken (Figs. 1 to 5, 7). Secondly, the turbine was monitored during one week with additionally performing some parameter variation studies. These included wind speed, generator activation procedures and activating/deactivating additional cable fixations (Fig. 1 and 6). It was found that the generator activation point coincided exactly with the rotor rotational frequency being identical to the system fundamental natural frequency. Adding of a tuned myss damper at the mast top solved the problem (Fig. 8).

 ev 600 Foto Ansicht total neu IMG_0211 mit Text e Kopie  ev600_1 Plan Ansicht Kopie
Fig.1: Wind Turbine envergate ev600 on the Migros Gossau Building roof. Left: Foto with the measurement levels indicated. Right: Drawing, dimensions: mm. Here, the foundation is indicated schematically only.


Fig. 2: Degrees of Freedom used for the modal analysis. Blue: references, green: rovers; two rovers per setup. Fig. 3: Sensors mounted at the mast and foundation. Magnetic support, additionally fixed with a textile strap at the mast.

Identical technical parameters were used for the modal analysis under ambient excitation and monitoring: LMS Pimento frontend, sampling rate sR = 200 Hz, length of the time window T = 30 minutes. The two differences remaining were that the sensors at measurement level 7 had to be dismounted and all six measurement points at the foundation were simultaneously instrumented for monitoring. Experimental modal analysis under ambient excitation using EFDD procedures allowed identification of 20 natural modes of the turbine in the frequency band f = 2.3…91.3 Hz including frequency, mode shape and damping. Figure 4 shows two EFDD SVD diagrams. As an illustration, Figure 5 presents the shape of the fundamental mode (mode 1a, f = 2.30 Hz, zeta = 0.62%; see also 2011 SGEB-DACH (in German)).

 ev600 EFDD SVD files 1 bis 3 dec 5 2K 1 bis 6 Hz mod
 Fig. 4: EFDD SDV-Diagram, f = 0…6 Hz.


ev600 mode 1a 2.302 Hz
Fig. 5: Mode m1a, f = 2.30 Hz, zeta = 0.62%.

The acceleration amplitudes observed during monitoring were about a = 0.01 m/s2 for weak, about a = 1.5 m/s2 for medium and a > 4 m/s2 for strong wind conditions. The safety switch disconnecting the generator and stopping the rotor was activated at about a = 2…3 m/s2. As an example, Figure 6 shows an acceleration signal measured at the mast top and the respective PSD-frequency spectrum.

 Fig. 6: Acceleration at the mast top with strong wind conditions: time signal and PSD-spectrum.

Having very well served for the test layout planning, the FE model generated before the tests, the so-called „baseline model“, was updated based on the modal analysis results (Fig. 7). Comparison showed that the number of modes identified with the experiment was significantly larger than that of the analytical modes. The reason for this was that the rotor was (had to be) neglected in the FE model. It was not possible to include the rotor in the experimental analysis because the resources available did not allow this (what sensor type? how to fix the sensor the CFK-blades?). However, next time we will include the rotor in the FE model although the experimental part will not be easier. This will definitely help to understand the turbine dynamic behavior.

Fig. 7: Mode Pair 3, FE model: f = 9.69 Hz (blue), experiment: f = 10.10 Hz (red), MAC = 0.702.

Monitoring of the turbine while being in operation had yielded that the generator was activated at a rotor frequency exactly coinciding with the turbine mast fundamental natural frequency (see Fig. 6). Subsequently to the investigation described here the designer of the wind turbine built an innovative tuned mass damper (Fig. 8). Although having been added to the Gossau turbine outside of the mast (!), this will be added to new turbines of this type inside of the mast at its top. The Gossau wind turbine works flawlessly now.


Fig. 8: Tuned Mass Damper (TMD) designed by envergate ag.