Simulation And Experiment Research On Detecting Approaches Of Harmonic Currents For Active Power Filter

In recent years, current rectifiers, frequency converters, arc furnaces and other power electronic devices etc have been increasingly used in power network. With nonlinear, impactive, unbalanced characteristics, they generate lots of harmonics and cause distortion in power network current or voltage. Power networks are polluted by harmonics, which do harm to electrical equipments; cause devices superheat, generate vibration and noise. A more serious consequence is that devices are damaged. Harmonic pollution has been one of the nonnegligible problems in power supply systems and now catches people's eye.Passive filters and active power filters (APF) are the two primary approaches for eliminating harmonics. Comparing with passive filters, APF have many advantages, which are not easily affected by system impedance and can be used to dynamically compensate harmonic. APF will play an important role in eliminating harmonics. Researches on APF are focused on main circuit structure forms (series-type, shunt-type, series- and shunt-type), harmonic current detecting approaches and control strategies to compensate current etc. This paper mainly deals with harmonic current detecting approaches.An accurate and real-time detecting instantaneous distorted current is critical for APF to accurately compensate harmonic currents. There are several harmonic currents detecting approaches for APF, for example, method based on FFT, ip -iq approach based on theory of instantaneous reactive power, ij -iq approach based on d-q transform and approach based on ANN (artificial neuron network). Harmonic currents detecting approach based on FFT is complicated and in need of considerable calculation, which cannot satisfy the need of realtime. Both ip -iq approach and id -iq approach need low pass filters, as a result the accuracy of harmonic current detection is influenced by frequency drift. Moreover, they both can only be used in three-phase power systems. With adaptive ability, harmonic current detecting approach, based on ANN, can accurately and real-timely detect harmonic currents for power network. In this paper, the classical ANN approach is improved because of the introduction of inertial coefficient. Now the ANN can use a bigger rate of learning. With the introduction of former k times sample values, the improved ANN method can only used to detect harmonic currents. We now take digital circuit approaches to realize the algorithms, i.e. using microprocessor to calculate harmonic components. So as to overcome disadvantages of analogy circuits, forexample, instable circuit parameters, low calculation accuracy, inconvenient regulation and low integration level. On the base of digital simulation, an experimental laboratory prototype is designed. Simulates and experimental results are satisfactory. It has been proved that the improved ANN method is valid, correct and practical.