7 Results: Comparative Studies
7.2 Comparison of frequency control schemes
Table 7.1 Comparison of results obtained before and after introduction of wind
power.
frequency nadir [Hz]
max. RoCoF
[mHz/s] Recovery time
without wind power 49,8 -70,8 4 min 30 s
with 40 % wind power 49,68 -109,1 4 min 56 s
7.2 Comparison of frequency control schemes
In the following, it will be examined how much the introduced control methods can contribute to frequency quality. The proposed strategies are tested for 40 % wind penetration, again under a 500 MW fault at t = 300 s. In case a), VHIE and in case b), pitch-controlled deloading will be observed individually before in case c), a combination of both is applied. As VHIE reacts to the RoCoF and the pitch control scheme to frequency deviations, a combination is expected to enhance frequency behaviour in respect to maximum RoCoF and frequency nadir. Furthermore, the over-speeding strategy is compared in case d), which reacts to frequency and RoCoF and is therefore not combined. They will further be compared to the previously studied case e), without IBR frequency support (no FFR).
Figure 7.2 represents the frequency gradients obtained for each case. It can be seen that all four FFR methods result in an improvement of the frequency nadir.
VHIE alone however, is not able to reach a frequency nadir within the regulatory limit. All cases involving deloading strategies however, that are b), c) and d), reach a frequency nadir above 49,85 Hz. This is due to the higher overall energy output, as will be discussed later. The combination of pitching and VHIE however does not add the improvements of both methods, but delivers a frequency nadir slightly below that of the pitching case, as VHIE reduces the overall energy output. over-speeding reaches the highest performance in respect to frequency nadir, arresting the drop at 49,89 Hz.
As VHIE reacts to the RoCoF, cases a) and c) delay the frequency nadir by 9 and 13 seconds. over-speeding, which also comprises a RoCoF control loop, delays the frequency nadir by 5 seconds. This behaviour is also represented in the maximum RoCoF, represented in Table 7.2, which summarises the obtained results.
page 54 7 Results: Comparative Studies
200 250 300 350 400 450 500 550 600 650 700 750 800 49,65
49,7 49,75 49,8 49,85 49,9 49,95 50 50,05
time [s]
frequency[Hz]
a) VHIE b) pitching
c) pitching + VHIE d) overspeeding e) no FFR
Figure 7.2 Comparison of frequency gradients for different frequency control strate-gies.
Table 7.2 Comparison of results obtained by different frequency control strategies.
control scheme frequency nadir [Hz]
max. RoCoF
[mHz/s] Recovery time
a) VHIE control 49,72 -107,1 4 min 35 s
b) Pitch control 49,86 -119,2 8 min 40 s
c) VHIE + pitch 49,85 -106,6 8 min 40 s
d) overspeed control 49,89 -106,6 8 min 32 s
e) no FFR 49,68 -123,8 4 min 56 s
A major drawback visible in the frequency graphs of all cases that involve deloading, that are b), c), and d), is the prolongation of recovery time by a factor of 1,76 as compared to the case of no FFR (seeTable 7.2). These control schemes, following their implemented droop, gradually decrease the power output as the post nadir frequency rises. As the grid’s preexisting PCR follows a similar droop, frequency recovery is slower than was anticipated when the droop for the thermal power plants was chosen. A more flat droop, resulting in higher power outputs, may therefore be considered.
7.2 Comparison of frequency control schemes page 55 Figure 7.3represents the power gradients under VHIE control. An instant power increase at the time of the fault occurrence can be observed. This leads to a lower maximum RoCoF, as can be seen in Table 7.2. Historically, RoCoF withstand capability has not been a design criteria for grid or power plant assets, however, high RoCoF causes additional wear and tear and thereby shortened life time [66].
The ENTSO-E specifies requirements on withstand capability for different grid components [66]. Generation units for example are required to stay connected up to a RoCoF of ±2Hz/s, calculated as moving average over 500 ms. In its frequency stability evaluation [54] however, the ENTSO-E states that for the SgCE, the maximum RoCoF that can be handled successfully lies in the range of 0,5 to 1Hz/s. This, as well all other asset’s limitations, are not exceeded under implementation of any of the proposed control techniques.
200 250 300 350 400 450 500 550 600 650 700 750 800 0,8
0,85 0,9 0,95 1 1,05 1,1 1,15
time [s]
frequency[Hz]
PMPP Pa Pgen
Figure 7.3 Power gradients of a wind turbine under VHIE control.
Further, Figure 7.3 shows the expected control behaviour for VHIE control as introduced in chapter 6.7.1. Following the contingency, power output immediately spikes as a reaction to the deteriorated RoCoF. When power is extracted from the rotating mass, its slowdown causes PMPP to drop, thereby also reducing the overall electric power output (see eq. 6.70). As a result of the lower tip-speed ratio, efficiency decreases and a slight decline in aerodynamic power is observed. Once Pgen falls belowPMPP0, frequency support is deactivated and the reference power is set to a
page 56 7 Results: Comparative Studies value between PMPP and Pa. The PI-controller tracks the reference signal with a short overswing and at t = 392 s generator power goes below PMPP +Pdeact. At this point, FS is deactivated and the power reference is set to PMPP, leading to an accelerated regeneration of rotor speed, but also to a secondary frequency dip which will be elaborated on in chapter 7.2.1.
Figure 7.4 represents the power gradient for case b). It is proportional to the frequency gradient in Figure 7.2. A significant delay due to the reaction time of mechanical pitching of the blades can not be observed due to the high sensitivity of the turbine’s power factorcp to the pitch angle, as elaborated on in chapter 6.8.
The step-wise adaption of power output visible in the graph can be ascribed to the lookup-table approach chosen in chapter 6.8. As this table is recalculated for each time step according to the current tip-speed ratio, a finer discretisation of power would result in increased computational effort which is not justified by the expected enhancement of the results.
200 250 300 350 400 450 500 550 600 650 700 750 800 0,9
0,91 0,92 0,93 0,94 0,95 0,96 0,97
time [s]
power[pu]
Pgen
Figure 7.4 Power gradient of a wind turbine under pitch control.
Figure 7.5 shows the power gradients for the combined control system. Under this regime, the decrease of aerodynamic power due to a reduced tip-speed ratio is countered by the pitching of the blades, causing an overall increase inPa. Because of this, PMPP is recovered before the available inertia is used up, causing Pgen to
7.2 Comparison of frequency control schemes page 57 reach PMPP. In this case, FS is deactivated according to condition c) described in chapter 6.7.1. No SR is necessary, as the transition condition Pel ≤PMPP+ ∆Pdeact is immediately fulfilled. This way a secondary frequency dip, as is typical for all synthetic inertia approaches, is avoided.
200 250 300 350 400 450 500 550 600 650 700 750 800 0,84
0,86 0,88 0,9 0,92 0,94 0,96 0,98 1 1,02
time [s]
power[pu]
PMPP
Pa
Pgen
Figure 7.5 Power gradients of a wind turbine under combined VHIE and pitch control.
7.2.1 The Effect of speed recovery in VHIE
To evaluate the effectiveness of the applied SR scheme of the VHIE control, VHIE with and without SR scheme will be compared in regard to the appearance of a secondary frequency dip in this section. The resulting frequency graphs are given in Figure 7.6.
page 58 7 Results: Comparative Studies
200 250 300 350 400 450 500 550 600 650 700 750 800 49,4
49,5 49,6 49,7 49,8 49,9 50 50,1
time [s]
frequency[Hz]
controlled SR uncontrolled SR
Figure 7.6 Frequency comparison under VHIE control with and without controlled SR.
Without controlled speed recovery, the generator power plunges down to the MPPT curve when rotational power is exhausted. The implemented SR prevents this by setting the electric power output betweenPMPP and Pa, as described in chapter 6.7.
The reulting power gradients for both cases are compared inFigure 7.7. A secondary frequency dip generally appears, when VHIE support is deactivated. In the controlled SR case, a slight secondary frequency dip is observed att= 392 s, decreasing frequency by 0,01 Hz. This frequency dip correlates with the deactivation of controlled SR (see Figure 7.7, left). In the uncontrolled case, given that rotational energy is exhausted before the frequency nadir is reached, no first and second frequency dip can be distinguished. Rather, a severe degradation of of the frequency nadir from 49,81 Hz for controlled SR to 49,68 Hz for uncontrolled SR is observed. Furthermore, the strong excursion of the frequency causes a stronger reaction of the grid’s conventional PCR, causing oscillations in the frequency gradient that last for approximately 1 minute after the contingency. Comparison of the power gradients further shows, that the SR implementation prolongs the speed recovery time of the turbine by a factor of 2,3. Recovery time of the frequency however is shortened by 17 seconds, as given isTable 7.3.