Railway electrified traction is more and more widely used in modern railway transportation because of its advantages of idle speed, heavy load, saving primary energy and reducing environmental pollution. However, the electric traction load is a kind of single-phase high-power rectification load with great volatility, and it is one of the most serious harmonic sources in the current power system. In order to reduce the adverse impact of the traction load on the power system, it is necessary to install a filter at the exit of the traction arm of the electric railway traction station to absorb the harmonic current generated by it.
Conventional passive filters cannot meet the filtering requirements of electric traction loads due to their own defects. The active power filter (APF), which has developed rapidly in recent years, has received more and more attention and application due to its many advantages. APF is an active compensation device. It calculates the harmonic current of the load through the detection link, and then controls the output of the main circuit to provide a compensation current that is opposite to the harmonic current of the load to eliminate the harmonics. Among them, the real-time detection of harmonic current is undoubtedly an important part of ensuring APF compensation characteristics.
For the harmonic current detection, the current application is based on the instantaneous reactive power theory detection method and the dqO system based generalized instantaneous reactive power detection method. Recently, the artificial neural network based adaptive harmonic detection is proposed. Method, etc. They mainly have two problems: First, due to the time delay of sampling and calculation, real real-time detection cannot be realized; second, the traction load changes rapidly, which is a strong time-varying load. The first two detection methods all contain low-pass filters, which need to conform to the actual comprehensive training samples. There is a significant time lag between the transient processes of current changes, and they cannot respond to the rapid changes of harmonic currents.
For the APF to do real-time compensation, it is necessary to know the harmonic current of the next switching moment in advance, but the next moment has not yet arrived, and the current cannot be detected in practice. Based on this, this paper proposes a predictive harmonic current detection method, which can accurately obtain the harmonic current in the future moment through the prediction link, and can track the rapid change of the load, and realize real-time control when the next moment arrives. No beat control.
2 Principle of Harmonic Current Detection in Single-Phase Circuits The three-phase circuit harmonic current detection method based on instantaneous reactive power theory is investigated. It is found that the detected three-phase signals are always changed into two-phase signals in mutually perpendicular coordinate systems. Then further calculations. For the single-phase circuit, the above method is simplified, and only one phase current and the actual current lag TV4 (here, the power frequency period) are reconstructed, and the hypothetical two-phase coordinate system signal can be directly formed.
Let the single-phase circuit current instantaneous value be more than one, sina>r and costwi are the standard sinusoidal signals obtained by the voltage signal through zero-crossing synchronization and phase-locked loop. "-" above the variable
Indicates the DC component, "" represents the AC component, and ip and Q represent the active and reactive DC components of the current, respectively. The fundamental current is obtained by performing an inverse transformation corresponding to the equation for iP and L. with…
f. Subtracting the fundamental component from the full current to obtain the harmonic current component ish=K-Ki (7) The single-phase circuit harmonic current detection principle based on the above method is as shown. Among them, the factory "is lagging / 4 links, C22 and C3 as shown in equation (8) 3 predictive harmonic current detection algorithm This paper proposes a predictive harmonic current detection algorithm to solve the aforementioned time delay problem to achieve fast Tracking, its implementation principle is as shown, where ia is the load current detected by the sensor, and PM is a prediction link.
Predictive type single-phase circuit harmonic current detection principle 3.1 Load current gradation case The sample point sequence of +1 point in the k-cycle is predicted to consist of two parts: hU fixed component, i2U+l) is the disturbance component.
For the fixed component, when the load current is gradual, considering the periodic characteristics, this paper uses the linear extrapolation algorithm, and for the disturbance component, the change is closely related to the previous sampling points. To improve the calculation accuracy, the second-order extrapolation is used. law.
3.2 The sudden change of load current According to the running condition of the locomotive and the number of locomotives on the power supply arm, the magnitude of the traction load often changes drastically. The harmonic current component at this time also changes drastically, and some corrections are needed for its prediction algorithm. The calculations show that when the total current of the electric locomotive changes, the fundamental component of the electric locomotive also has a similar change. When the electric locomotive passes through the partition kiosk, for a power supply arm of the traction power supply system, a locomotive is added or reduced. In the locomotive, the traction arm load current exhibits a step change characteristic. At this time, the following steps can be used to process the above steps.
After this processing, the predicted harmonic current can quickly track the load change and solve the hysteresis effect of the general low-pass filter.
In the actual calculation, firstly, they all contain 3, 5, 7, 9, 11 harmonic components, and b and b represent the harmonic current results detected by conventional algorithms and prediction algorithms under different variations. It can be seen that when the load current changes gradually, the harmonic current obtained by the prediction algorithm is basically consistent with the theoretical harmonic current, while the conventional algorithm has a large error-changing load current model containing a drastically changing load current model ( The current is increased by a factor of 1 in 1/4 cycle. It shows that when the load current changes drastically, the harmonic current obtained by the conventional algorithm has about one period of time lag, and the prediction algorithm has basically no time lag.
5 Conclusions This paper presents a predictive harmonic current detection method. The condition can still give accurate prediction values, and overcome the time lag problem existing in other methods, which lays a foundation for the active filter system to filter harmonic current. This harmonic detection method can be used for harmonic current detection such as general rectification load and electric traction load.
Men'S Casual Sportswear,Sportswear Casual Wear,Sportswear Casual,Casual Sportswear Brands
SUZHOU HENGRUN IMP. & EXP. CORP.,LTD , https://www.hengrunsz.com