CN105403771A - Improved adaptive principle harmonic detection method - Google Patents
Improved adaptive principle harmonic detection method Download PDFInfo
- Publication number
- CN105403771A CN105403771A CN201510587588.XA CN201510587588A CN105403771A CN 105403771 A CN105403771 A CN 105403771A CN 201510587588 A CN201510587588 A CN 201510587588A CN 105403771 A CN105403771 A CN 105403771A
- Authority
- CN
- China
- Prior art keywords
- harmonic
- adaptive
- detection method
- current
- algorithm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses an improved adaptive principle harmonic detection method. An active power filter employing the method is formed by a sampling circuit, harmonic current detection, control algorithm realization, trigger pulse formation and protection, an IGBT driving circuit, and a main circuit. According to the method, orthogonal signals i[alpha] and i[beta] are obtained via Clark coordinate transformation of three-phase load current ia, ib, and ic; the orthogonal signals i[alpha] and i[beta] are filtered via a low-pass filter and regarded as inputs x1 (n) and x2 (n) of the adaptive harmonic detection method; y (n) is obtained via an adaptive algorithm of x1 (n) and x2 (n); a harmonic command current e (n) is obtained by subtracting y (n) from the load current iL (n); and e (n) is fed back to an input terminal, and a new y (n) is obtained via the adaptive algorithm of e (n) and input signals. According to the conventional adaptive harmonic detection method, two inputs x1 (n) and x2 (n) are built via the locking of power grid phases by a phase-locked loop. According to the detection method, good dynamic response speed of harmonic detection can be obtained, the filtering precision is improved, and the anti-interference capability of the system is enhanced.
Description
Technical field
The invention discloses a kind of adaptive principle harmonic detecting method of improvement, belong to electric and electronic technical field.
Background technology
Along with the development of science and technology, non-linear electrical equipment puts into operation in a large number in electrical network, and Harmonic in Power System content is more and more serious, even threatens the safe operation of electrical network.Nowadays administer mains by harmonics and mainly adopt passive power filter (LC) and Active Power Filter-APF (APF).Active Power Filter-APF is because can compensate each harmonic, and the harmonic wave that all can change frequency and amplitude realizes dynamic compensation and the resonance problems that can overcome LC wave filter intrinsic is widely used.The key property of Active Power Filter-APF depends on the method extracting harmonic wave instruction current, produces the control method of compensation harmonic electric current, and these three aspects of the dynamic perfromance of whole system.
Based on self-adaprive predictive control adaptive neural network due to can be accurately real-time extraction harmonic wave, and it is dark to have zero point, and the advantages such as error is little, obtain and pay attention to more and more widely.The reference input of adaptive neural network is two orthogonal signal, and current document is all generally with the synchronous standard sine signal of line voltage, and the cosine signal obtained after 900 phase shifts is as reference input.But this method need use a phase-locked loop module, not only can reduce the real-time of harmonic detecting, the time of adaptive algorithm convergence also can be extended.
Summary of the invention
Technical matters to be solved by this invention is the deficiency for above-mentioned background technology, provides a kind of adaptive principle harmonic detecting method of improvement.
The present invention adopts following technical scheme for achieving the above object:
An adaptive principle harmonic detecting method for improvement, specifically comprises the steps:
Step 1, by threephase load current i
la, i
lb, i
lcthrough the orthogonal signal i that Clark coordinate transform obtains
l α, i
l β;
Step 2, i
l α, i
l βas the input x of adaptive neural network after wave digital lowpass filter filtering
1(n), x
2(n);
Step 3, x
1(n), x
2n () obtains y (n) through adaptive algorithm;
Step 4, load current i
ln () deducts y (n) and obtains harmonic wave instruction current e (n);
Clark coordinate in described step 1 is transformed to:
Wave digital lowpass filter in described step 2 is three rank wave digital lowpass filters.
The present invention adopts technique scheme, has following beneficial effect: omit phaselocked loop link, improves dynamic responding speed, improves filtering accuracy, strengthens antijamming capability.
Accompanying drawing explanation
Fig. 1 is adaptive neural network schematic diagram of the present invention.
Fig. 2 is second order self-adaptive harmonic detecting schematic diagram of the present invention.
Fig. 3 is Butterworth tri-rank of the present invention LPF frequency response charts.
Fig. 4 is the process flow diagram of LMS adaptive algorithm of the present invention.
Fig. 5 is that the present invention compensates front A phase power network current waveform.
Fig. 6 is that the present invention compensates rear A phase power network current waveform.
Fig. 7 is the fundamental current waveform that innovatory algorithm of the present invention detects.
Fig. 8 is the harmonic current waveforms that innovatory algorithm of the present invention detects.
Fig. 9 is innovatory algorithm input signal x of the present invention
1(n).
Figure 10 is the fundamental current waveform detected before the present invention improves.
Figure 11 is the harmonic current waveforms detected before the present invention improves.
Figure 12 is the improvement Harmonic detection control block diagram that the present invention is based on LMS algorithm.
Figure 13 is that during the present invention tests, APF drops into front line voltage and current waveform.
Figure 14 is the harmonic wave forms detected after innovatory algorithm during the present invention tests.
Embodiment
Be described in detail below in conjunction with the technical scheme of accompanying drawing to invention:
As shown in Figure 1, n (n) is the first-harmonic composition of load current, and S (n) is the harmonic components of load current.X (n) is synchronous standard sine signal with line voltage, and e (n) is the harmonic wave instruction current expected.Its ultimate principle is as follows:
e
2(n)＝s
2(n)+[n(n)-y(n)]
2
(2)
+2s(n)[n(n)-y(n)]
E (n)=s (n)+[n (n)-y (n)] (3) output error square be:
e
2(n)＝s
2(n)+[n(n)-y(n)]
2
(4)
+2s(n)[n(n)-y(n)]
Here s (n) is uncorrelated with y (n), n (n), so have:
E[e
2(n)]＝E[s
2(n)]+E{[n(n)-y(n)]
2}
+E{2s(n)[n(n)-y(n)]}
＝E[s
2(n)]+E{[n(n)-y(n)]
2}
Adopt least fibre method (LeastMeanSquare, LMS), regulate the parameter of sef-adapting filter to make E [e
2(n)]=E
min[e
2(n)] time, have:
E{[n(n)-y(n)]
2＝E
min{[n(n)-y(n)]
2(5)
Now, y (n) and n (n) difference are minimum, and namely sef-adapting filter achieves and well to estimate fundamental current and follow the tracks of.
With one of adaptive algorithm be most widely used, least fibre method is example, lowest mean square LMS method, based on steepest descent method, constantly updates weights coefficient according to the gradient vector of mean square of error value between wanted signal and output signal in an iterative process, minimum to meet square error.In sef-adapting filter, according to least-mean-square-error criterion, best weight w should make sef-adapting filter performance function:
f(w)＝ξ＝E{e
2(n)}(6)
For minimum.Above formula is called square error performance function, and its image is the parabola of an opening upwards.When application steepest descent method finds optimal weight vector on hyperparaboloid, first to determine that in functional image, certain point, as starting point, is supposed it is w (0), then deducted a vector along its gradient direction and obtain second some w (1).According to said method carry out iteration successively later, progressively search the minimum point on square error performance function image, the w of this place's correspondence is exactly optimum solution, and now sef-adapting filter is operated in optimum condition.
The iterative formula of LMS adaptive algorithm is as follows:
e(n)＝d(n)-y(n)＝d(n)-X
T(n)W(n)(7)
W(n+1)＝W(n)+μe(n)X(n)(8)
Here, X (n) represents input signal vector, and W (n) represents weights coefficient vector, and it is very large to LMS convergence of algorithm process influence that μ represents step-length factor mu, and μ is sufficiently little, to ensure convergence.When ensureing convergence, μ is larger, and convergence is faster, but when μ is too large, the process of transition will be vibrated.General, the condition of μ demand fulfillment is as follows:
Wherein, λ
maxit is the eigenvalue of maximum of X (n) autocorrelation matrix.
Adapt to harmonic detecting method requirement, input reference signal is relevant to first-harmonic, uncorrelated with harmonic wave.In order to be met the input signal of condition, traditional algorithm is all generally obtain electric network voltage phase by phaselocked loop, constructs one group and the synchronous cosine and sine signal of line voltage.But phaselocked loop inner structure is complicated, operand is large, and will use integral controller, will increase the harmonic detecting time like this, reduces the dynamic response of APF.The input signal of second order self-adaptive wave filter is as shown in Figure 2 two orthogonal signal x
1(n) and x
2(n), x
1n () is the standard sine signal synchronous with line voltage obtained with phaselocked loop, x in traditional algorithm
2n () is x
1n standard cosine signal that () obtains after 90 ° of phase shifts.W
1and w
2represent reference-input signal x respectively
1(n) and x
2the weights of (n).
The present invention is directly by threephase load current i
la, i
lb, i
lcthrough the orthogonal signal i that Clark coordinate transform obtains
l α, i
l βshown in (10), orthogonal signal i
l α, i
l βas the input of adaptive neural network after wave digital lowpass filter filtering, as shown in Figure 2.
In three-phase symmetrical circuit, only containing 6k ± 1 subharmonic in load current, select three rank Butterworth low-pass filters herein, cutoff frequency is 100Hz, thus ensures in the output signal of harmonic detecting method only containing first-harmonic.Butterworth third-order low-pass filter frequency response chart as shown in Figure 3.
In order to verify the adaptive neural network validity of improvement, having built MATLAB/Simulink model herein and having carried out discretize emulation, its program realizes as shown in Figure 5.Simulation parameter arrange and result as follows:
Table 1 is based on the improvement Harmonic detection simulation parameter of LMS algorithm
Before and after compensating, the waveform of A phase power network current respectively as shown in Figure 6 and Figure 7, load changing when 0.2 second, compensate through Active Power Filter-APF, power network current is close sinusoidal in the time less than one-period, before 0.2s, current distortion rate is 1.92%, and load changing after-current aberration rate is 3.85%.
Fig. 8 is the output fundamental current waveform that the adaptive neural network of improvement is corresponding, and Fig. 9 is the output harmonic wave current waveform that the adaptive neural network of improvement is corresponding.Figure 10 is one of the input signal that innovatory algorithm is corresponding x
1(n).As can be seen from the figure, load changing during 0.2s, the time about one-period that now sef-adapting filter detects just starts to tend towards stability.
In order to compare the transient response speed of harmonic detecting before and after innovatory algorithm, when ensureing other parameter constants, reference signal x (n) and x
90 °n () is obtained through phaselocked loop by line voltage.As can be seen from the results, before not improving harmonic detecting method, the fundamental current of output just tends towards stability as shown in FIG. 11 and 12 more than two primitive periods.
The present invention utilizes existing experiment porch to carry out experimental verification to the validity of innovatory algorithm.
Table 2 experiment parameter
The present invention, for the adaptive neural network based on LMS algorithm, has set forth its ultimate principle and programming realization process respectively.Improve adaptive neural network, can find out that the method after improving significantly improves the dynamic responding speed of harmonic detecting by simulation result, last experimental result also demonstrates the validity of this innovatory algorithm.
Finally, note also that, above embodiment, only for illustration of the present invention, is not limitation of the present invention.The content described that do not elaborate in instructions of the present invention belongs to the known prior art of professional and technical personnel in the field.
Claims (3)
1. the adaptive principle harmonic detecting method improved, is characterized in that, comprise the following steps:
Step 1, by threephase load current i
la, i
lb, i
lcthrough the orthogonal signal i that Clark coordinate transform obtains
l α, i
l β;
Step 2, i
l α, i
l βas the input x of adaptive neural network after wave digital lowpass filter filtering
1(n), x
2(n);
Step 3, x
1(n), x
2n () obtains y (n) through adaptive algorithm;
Step 4, load current i
ln () deducts y (n) and obtains harmonic wave instruction current e (n).
2. the adaptive principle harmonic detecting method of a kind of improvement according to claim 1, is characterized in that, the Clark coordinate in described step 1 is transformed to:
。
3. the adaptive principle harmonic detecting method of a kind of improvement according to claim 1, is characterized in that, the wave digital lowpass filter in described step 2 is three rank wave digital lowpass filters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510587588.XA CN105403771A (en) | 2015-09-15 | 2015-09-15 | Improved adaptive principle harmonic detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510587588.XA CN105403771A (en) | 2015-09-15 | 2015-09-15 | Improved adaptive principle harmonic detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105403771A true CN105403771A (en) | 2016-03-16 |
Family
ID=55469377
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510587588.XA Pending CN105403771A (en) | 2015-09-15 | 2015-09-15 | Improved adaptive principle harmonic detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105403771A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106959385A (en) * | 2017-04-21 | 2017-07-18 | 燕山大学 | The Harmonic currents detection method of phase is locked during unbalanced source voltage based on two frequencys multiplication |
| CN108152585A (en) * | 2017-12-27 | 2018-06-12 | 江苏中科君芯科技有限公司 | Adaptive neural network and detection circuit based on neural network |
| CN108226637A (en) * | 2017-01-04 | 2018-06-29 | 中国矿业大学(北京) | A kind of any order component detection method with frequency variation adaptability |
| CN109633272A (en) * | 2019-01-22 | 2019-04-16 | 燕山大学 | A kind of harmonic detecting system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101187682A (en) * | 2007-11-26 | 2008-05-28 | 天津理工大学 | On-line self-adaptive harmonic current detection system |
| CN101726648A (en) * | 2009-12-18 | 2010-06-09 | 国网电力科学研究院武汉南瑞有限责任公司 | Neuron adaptive harmonic current detection circuit based on low pass filter |
| CN103336173A (en) * | 2013-01-29 | 2013-10-02 | 上海海维工业控制有限公司 | Genetic algorithm based self-adaption harmonic detection method |
| CN103424621A (en) * | 2013-08-20 | 2013-12-04 | 江苏大学 | Artificial neural network detecting method of harmonic current |
-
2015
- 2015-09-15 CN CN201510587588.XA patent/CN105403771A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101187682A (en) * | 2007-11-26 | 2008-05-28 | 天津理工大学 | On-line self-adaptive harmonic current detection system |
| CN101726648A (en) * | 2009-12-18 | 2010-06-09 | 国网电力科学研究院武汉南瑞有限责任公司 | Neuron adaptive harmonic current detection circuit based on low pass filter |
| CN103336173A (en) * | 2013-01-29 | 2013-10-02 | 上海海维工业控制有限公司 | Genetic algorithm based self-adaption harmonic detection method |
| CN103424621A (en) * | 2013-08-20 | 2013-12-04 | 江苏大学 | Artificial neural network detecting method of harmonic current |
Non-Patent Citations (1)
| Title |
|---|
| 刘艳英: "基于自适应理论的模块化并联型有源滤波器研究", 《中国优秀硕士学位论文全文数据库 工程科技II辑》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108226637A (en) * | 2017-01-04 | 2018-06-29 | 中国矿业大学(北京) | A kind of any order component detection method with frequency variation adaptability |
| CN106959385A (en) * | 2017-04-21 | 2017-07-18 | 燕山大学 | The Harmonic currents detection method of phase is locked during unbalanced source voltage based on two frequencys multiplication |
| CN108152585A (en) * | 2017-12-27 | 2018-06-12 | 江苏中科君芯科技有限公司 | Adaptive neural network and detection circuit based on neural network |
| CN109633272A (en) * | 2019-01-22 | 2019-04-16 | 燕山大学 | A kind of harmonic detecting system |
| CN109633272B (en) * | 2019-01-22 | 2020-04-07 | 燕山大学 | Harmonic detection system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108023352B (en) | Power grid high-frequency impedance remodeling device and method for inhibiting distributed generation resonance | |
| CN103560690B (en) | Harmonic wave damping control method for one-phase LCL type grid-connected inverter | |
| CN102118028B (en) | Method for suppressing and controlling current harmonics of three-phase LCL (Lower Control Limit) type grid-connected inverter | |
| CN103472282B (en) | A kind of FBD harmonic current detecting method based on adaptive principle | |
| CN104135003B (en) | APF control method based on active disturbance rejection and repetitive control | |
| CN110429603B (en) | Six-switch seven-level active power filter and compensation method | |
| CN109638881B (en) | Virtual synchronization method and system for energy storage inverter with power grid strength self-adaptive optimization | |
| CN105403771A (en) | Improved adaptive principle harmonic detection method | |
| CN103595280A (en) | Deadbeat control method of solar power generating system inverter under condition of voltage unbalance | |
| CN110943632A (en) | Energy storage converter virtual inertia control method based on cascade generalized integrator | |
| CN100461580C (en) | A harmonic current prediction method for compensating control delay | |
| Fei et al. | A novel sliding mode control technique for indirect current controlled active power filter | |
| CN112909919A (en) | Neutral point potential balancing method and system for two-stage three-level three-phase four-wire system energy storage converter | |
| CN102332721B (en) | Method for predicting harmonic current of hybrid active power filter based on optimal linear prediction theory | |
| CN112000018B (en) | Robust fault-tolerant control module, method and system based on residual generator | |
| Zhou et al. | Control research of NPC three level high-power grid connected inverter based on multi sampling | |
| CN113014250A (en) | Phase-locked loop capable of eliminating direct current offset voltage and phase-locked control method thereof | |
| CN107968406A (en) | A kind of adaptive repetitive control of Active Power Filter-APF anti-saturation frequency | |
| CN112688338A (en) | UPQC power quality compensation control method based on frequency-locked loop steady-state linear Kalman filtering | |
| CN115912489A (en) | LMS-SOGI three-phase-locked loop design method and system suitable for non-ideal power grid | |
| Wang et al. | Grid Impedance Detection Based On Complex Coefficient Filter and Full-order Capacitor Current Observer for Three-phase Grid-connected Inverters | |
| CN104578070A (en) | Double-voltage closed-loop-control harmonic current detection method and system | |
| Laggoun et al. | Dual Virtual Flux-based Direct Power Control for rectifier under harmonically distorted voltage condition | |
| CN113013915B (en) | Energy storage converter virtual inertia realization method based on embedded generalized integrator | |
| CN114566965A (en) | Single-phase UPQC voltage harmonic control method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20160316 |