PhD defence by Shan He
08.08.2022 kl. 13.00 - 16.00
Shan He, AAU Energy, will defend the thesis "Multi-Sampled Current Control of Grid-Connected Converters"
Multi-Sampled Current Control of Grid-Connected Converters
Professor Frede Blaabjerg
Associate Professor Dao Zhou
Professor Xiongfei Wang
Associate Professor Florin Iov, Aalborg University, Denmark (Chairman)
Professor Yaow-Ming Chen, National Taiwan University
Norbert Hanigovszki, Danfoss Drives
Due to the increasing performance and decreasing price of micro-controllers in recent years, applying a high sampling frequency becomes more feasible in the modern control that is known as multi-sampling technology. Besides the average voltage/current value, the switching ripple can also be sampled when using multi-sampling, which is the main difference compared with the regular single/double-sampling. Nevertheless, how to fully employ the potential benefits from multi-sampling to enhance the stability and the reliability of grid-connected converters is still missed.
First, multi-sampling control delay is inversely proportional to the sampling rate, which can make it close to the analog control if the sampling rate is high enough. However, based on the prior art, the mechanism of multi-sampling pulse width modulation (PWM) is still not fully investigated, and the effect of the sampled switching ripple on the control is also not clear. Moreover, the multi-sampling PWM is only applied in a few power electronic converters, and the potentiality of multi-sampling on other effective converters is still unknown. Thus, further analysis on the multi-sampling PWM principle and its feasibility is necessary.
Second, with the large-scale penetration of renewable generation, the grid admittance seen from the point of common coupling varies in a wide range, which poses a significant challenge to the harmonic stability of the inverter-grid system. Since the control delay plays an important role on the stability, how to use the multi-sampling PWM with less control delay to enhance the stability and even to achieve the passivity is worth to be investigated.
Third, apart from the reduced control delay and the improved stability from the multi-sampling control, some new states and parameters can be estimated based on the multi-sampled voltage/current data. The main principle is to utilize the sampled switching ripple to unleash more possibilities to save the cost and enhance the reliability. Therefore, the advanced condition monitoring methods are of importance to fully take advantage of multi-sampling technology.
In order to cope with the aforementioned issues, this Ph.D. project proposed a graphical analyzing method based on the voltage-second balance principle, where the relationship between the multi-sampling PWM and the double-sampling PWM is revealed. Then the low-order aliasing in the grid-side current can be explained when the switching harmonics are introduced in the control loop. In order to suppress the aliasing, an improved anti-aliasing filter is proposed to remove the sampled switching harmonics and to maintain the advantage of the phase boost using multi-sampling.
For a three-phase two-level inverter, the stability can be enhanced using single-loop multi-sampled control with inverter-side current feedback. However, the passivity still cannot be achieved due to the delay of the anti-aliasing filter in the feedback path. An active damping method using capacitor voltage feedforward is proposed to achieve passivity in the whole frequency range. On the other hand, when using the multi-sampled grid-side current control, a similar robust damping method using capacitor voltage feedforward and capacitor current active feedback is proposed. Further, the proposed methods are extended to the single-phase H-bridge inverter by doubling the sampling rate based on the equivalent switching frequency.
Afterwards, based on the multi-sampling equivalent delay concept, an enhanced real-time current control is proposed where the anti-aliasing filter is not required. Only the state at the peak/valley and the middle point of the carrier is sampled for the control, and the control delay can be optimized to one quarter of switching period which is the stability boundary based on the passivity theory. Moreover, compared with the traditional real-time-update double-sampling control, the duty cycle limitation is also removed.
Then, a multi-sampling control strategy is proposed for the interleaved three-phase converter with an L filter. The sampling rate selection and the controller design in the main current control loop and the circulating current control loop are discussed, respectively. In addition, a simple grid impedance estimation method is proposed by multi-sampling the voltage at the point of common coupling.
Besides the benefits on the stability enhancement, a grid voltage sensorless control method is proposed using the linear regression of the multi-sampled current data, which can help to reduce the cost and enhance the reliability under a voltage sensor fault. In order to address the transients during start-up, a soft start-up method is proposed where the inverter is regarded as a boost converter and the grid information can be known in advance based on multi-sampling. On the other hand, by four-sampling the inverter-side current, the grid impedance can be estimated for a two-cell three-phase interleaved inverter under an inductive grid.
The findings of this Ph.D. project can promote the application of multi-sampling technology on grid-connected converters.
THE DEFENCE will be IN ENGLISH - all are welcome.
Pontoppidanstræde 111, Room 1.177 / Online