Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Forming method for neural network inverse controller of active power filter

A neural network inverse, power filter technology, applied in active power filtering, AC network to reduce harmonics/ripples, harmonic reduction devices, etc., can solve the problems of control system performance degradation, decoupling condition destruction, etc. Achieve the effect of improving dynamic response performance, improving power quality, and improving the ability to cope with system parameter changes and external disturbances

Active Publication Date: 2014-05-28
江苏集萃中以科技产业发展有限公司
View PDF1 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The traditional analytical inverse decoupling control method relies heavily on the precise mathematical model of the system. When the system model parameters change, the decoupling conditions of the system will be destroyed, resulting in a decline in the performance of the control system, making the active power The analytical inverse control method of the filter system has a "bottleneck" in practical engineering applications

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Forming method for neural network inverse controller of active power filter
  • Forming method for neural network inverse controller of active power filter
  • Forming method for neural network inverse controller of active power filter

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0020] The embodiment of the present invention is: first, the inverter of the main circuit of the active power filter and the AC side inductance are taken as a whole to form the controlled object, and the controlled object is equivalent to two first-order differential equations under dq rotating coordinates model, the relative order of the vector of the system is {1, 1}; then a static neural network with 4 input nodes and 2 output nodes plus 2 integrators is used to form the neural network inverse of the controlled object through a learning algorithm, where the static The weight coefficients of the neural network are determined by learning; then the neural network is inversely connected in front of the controlled object, and the neural network inverse and the controlled object form a pseudo-linear system composed of two independent first-order integral compensation current subsystems, so that the The control of a multivariable nonlinear coupling system is transformed into the c...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a forming method for a neural network inverse controller of an active power filter, and the method comprises the following steps of: taking an inverter and an alternating current-side inductor of a main circuit of the active power filter as a whole to be formed into a controlled object, wherein the controlled object takes two switch function control quantity of the main circuit as input, and takes two compensating current components as output; according to an inverse system which corresponds to the controlled object, forming into neural network inversion by a static neural network and an integrator; equalizing the neural network inversion into a pseudolinear system consisting of two independent first-order integration type compensating current subsystems before the neural network inversion is connected with the controlled object in series; respectively designing two current controllers to the two independent first-order integration type compensating current subsystems in the pseudolinear system to form into a linear closed-loop controller; and connecting the linear closed-loop controller with the neural network inversion in series to be formed into a neural network inversion system method controller, so that an active power filter model can be nonlinearly controlled in a decoupling way at high performance, and the power quality of a power system can be improved.

Description

technical field [0001] The invention relates to a construction method of a neural network inverse controller of an active power filter, which is suitable for nonlinear decoupling control of an active filter in a power system and belongs to the technical field of power quality in a power system. Background technique [0002] With the increasingly higher requirements for power quality in the information age, harmonic suppression and reactive power compensation of power systems have become a hot spot in the field of power system research. The use of active power filters for harmonic and reactive power compensation is a future development trend. As a power electronic device that dynamically suppresses harmonics, active power filters can compensate for harmonics that change in frequency and size. . [0003] Since the model of the active power filter in the dq rotating coordinate system has nonlinear and parameter coupling problems, the inverse system method in the nonlinear syst...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): H02J3/01
CPCY02E40/22Y02E40/40Y02E40/20
Inventor 刘国海陈兆岭杨辰星
Owner 江苏集萃中以科技产业发展有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products