CN109217651B - A kind of APFC control system of online compensation control rate - Google Patents
A kind of APFC control system of online compensation control rate Download PDFInfo
- Publication number
- CN109217651B CN109217651B CN201811167864.7A CN201811167864A CN109217651B CN 109217651 B CN109217651 B CN 109217651B CN 201811167864 A CN201811167864 A CN 201811167864A CN 109217651 B CN109217651 B CN 109217651B
- Authority
- CN
- China
- Prior art keywords
- sliding mode
- neural network
- module
- control
- apfc
- 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.)
- Active
Links
- 239000002245 particle Substances 0.000 claims abstract description 58
- 238000013528 artificial neural network Methods 0.000 claims abstract description 57
- 238000000034 method Methods 0.000 claims abstract description 10
- 230000001939 inductive effect Effects 0.000 claims abstract description 8
- 238000005457 optimization Methods 0.000 claims description 20
- 230000006870 function Effects 0.000 claims description 13
- 238000012549 training Methods 0.000 claims description 10
- 239000013598 vector Substances 0.000 claims description 7
- 238000012937 correction Methods 0.000 claims description 5
- 230000003247 decreasing effect Effects 0.000 claims description 3
- 230000001447 compensatory effect Effects 0.000 claims 1
- 230000008569 process Effects 0.000 abstract description 5
- 238000009415 formwork Methods 0.000 abstract 5
- 238000001914 filtration Methods 0.000 abstract 1
- 230000006872 improvement Effects 0.000 description 7
- 239000003990 capacitor Substances 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 2
- 230000033001 locomotion Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 241000764238 Isis Species 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02M—APPARATUS FOR CONVERSION BETWEEN AC AND AC, BETWEEN AC AND DC, OR BETWEEN DC AND DC, AND FOR USE WITH MAINS OR SIMILAR POWER SUPPLY SYSTEMS; CONVERSION OF DC OR AC INPUT POWER INTO SURGE OUTPUT POWER; CONTROL OR REGULATION THEREOF
- H02M1/00—Details of apparatus for conversion
- H02M1/42—Circuits or arrangements for compensating for or adjusting power factor in converters or inverters
- H02M1/4208—Arrangements for improving power factor of AC input
- H02M1/4225—Arrangements for improving power factor of AC input using a non-isolated boost converter
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02B—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO BUILDINGS, e.g. HOUSING, HOUSE APPLIANCES OR RELATED END-USER APPLICATIONS
- Y02B70/00—Technologies for an efficient end-user side electric power management and consumption
- Y02B70/10—Technologies improving the efficiency by using switched-mode power supplies [SMPS], i.e. efficient power electronics conversion e.g. power factor correction or reduction of losses in power supplies or efficient standby modes
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a kind of APFC control systems of online compensation control rate, including Boost APFC main circuit, sliding formwork control circuit, Boost APFC main circuit will be converted to high voltage direct current after AC power source rectifying and wave-filtering, sliding formwork control circuit is controlled from the collected inductive current of Boost APFC main circuit and output voltage, realize PFC, and the output voltage of smooth steady is obtained, the sliding formwork control circuit includes outer voltage PI control module, sliding mode controller, PWM comparator and drive circuit module；Wherein the sliding mode controller includes sliding-mode surface module, sliding formwork control module, Elman neural network algorithm module and particle swarm algorithm module；The present invention efficiently solves existing DC-DC converter using traditional sliding formwork control, cause sliding-mode surface function level off to zero convergence rate it is slower, and sliding process deposits the problem of may being unable to complete in case of interferers in system.
Description
Technical Field
The invention relates to the technical field of power control equipment, in particular to an APFC control system for compensating a control rate on line.
Background
An AC/DC converter in the charging device of the electric automobile mostly adopts a boost active power factor correction (BoostAPFC) circuit, and after an alternating current input power supply is rectified and filtered, the input current contains a large amount of harmonic components, so that the input current generates distortion and has low power factor.
The APFC circuit comprises a nonlinear element, so that the traditional linear control method is difficult to achieve a satisfactory control effect, and the influence of system parameter disturbance is not considered. The Sliding Mode Control (SMC) can continuously adjust the control quantity to enable the system state to reach the set sliding mode surface and move along the track, can realize the effective control of a nonlinear system, and is successfully applied to a DC-DC converter. However, in the conventional SMC, the convergence speed of the sliding mode surface function approaching zero is slow, and the sliding process may not be completed in the presence of system interference.
Disclosure of Invention
The invention provides an APFC control system for compensating a control rate on line, which effectively solves the problems that the convergence speed of a sliding mode surface function approaching zero is slow and the sliding process can not be finished under the condition that the system is interfered because the traditional sliding mode control is adopted in the traditional DC-DC converter.
The invention is realized by the following technical scheme:
the APFC control system comprises a Boost APFC main circuit and a sliding mode control circuit, wherein the Boost APFC main circuit rectifies and filters an alternating current power supply and converts the rectified and filtered alternating current power supply into high-voltage direct current, and the sliding mode control circuit controls inductive current and output voltage acquired from the Boost APFC main circuit, so that power factor correction is realized, and smooth and stable output voltage is obtained.
The invention further adopts the technical improvement scheme that:
the sliding mode control circuit comprises a voltage outer ring PI control module, a sliding mode controller, a PWM comparator and a driving circuit module; the sliding mode controller comprises a sliding mode surface module, a sliding mode control module, an Elman neural network algorithm module and a particle swarm algorithm module;
the voltage outer ring PI control module is used for generating reference current and transmitting the reference current to the sliding mode surface module;
the sliding mode surface module is used for forming a sliding surface of the sliding mode controller;
the Elman neural network algorithm module is used for estimating the compensation control rate on line, and the particle swarm algorithm module is adopted to optimize the weight of the Elman neural network to obtain the optimized compensation control rate;
the sliding mode control module is used for receiving the sliding mode surface obtained by the sliding mode surface module, generating an equivalent control rate, adding the equivalent control rate and the optimized compensation control rate obtained by the Elman neural network algorithm module, and outputting a total sliding mode control rate;
the PWM comparator is used for comparing the total sliding mode control rate with a preset triangular carrier signal to generate a PWM signal and transmitting the PWM signal to the driving circuit module;
the driving circuit module is used for driving the power switch tube of the Boost APFC main circuit to be switched on and off.
The invention further adopts the technical improvement scheme that:
the PI control module comprises a PI controller and a multiplier, the PI controller is used for carrying out proportional integral control on the error between the direct current output voltage and the reference voltage so that the direct current output voltage tracks the reference voltage, and the multiplier is used for receiving the output quantity of the PI controller and multiplying the output quantity by the unit sine half wave of the power grid voltage to generate reference current.
The invention further adopts the technical improvement scheme that:
the sliding mode surface module is used for calculating the error between the reference current and the inductive current of the Boost APFC main circuit to form a sliding mode surface of the sliding mode controller.
The invention further adopts the technical improvement scheme that:
the particle swarm optimization module for optimizing the weight of the Elman neural network comprises the following steps:
step 1): initializing the structure of an Elman neural network, encoding the weight vectors from a parameter input layer to a hidden layer and the weight vectors from the hidden layer to an output layer as parameters of particles, randomly generating an initial particle population in a solution space, wherein each encoding string represents weight distribution, and initializing the scale, position and speed of a particle swarm;
step 2): carrying out sample training on the neural network;
step 3): calculating a fitness function value of each particle, thereby determining an individual extremum and a global extremum;
step 4): comparing the individual fitness of the particles with the fitness corresponding to the individual extreme value and the global extreme value respectively, and updating the individual extreme value and the global extreme value if the individual fitness is better than the global extreme value;
step 5): updating the inertia weight by adopting a mode of linearly decreasing the inertia weight, and performing iterative update on the speed and the position of the particle;
step 6): judging whether the optimization target is met or the maximum iteration number is reached, and if the optimization target meets the conditions, performing step 7); otherwise, obtaining a group of new particle swarms to generate a new Elman neural network, and returning to the step 2);
step 7): and decoding the global extremum of the particle group to serve as an optimized Elman neural network parameter, learning by the neural network, and finishing optimization.
The invention further adopts the technical improvement scheme that:
the method for obtaining the optimized compensation control rate by the Elman neural network algorithm module comprises the following steps:
step 21): acquiring an initial weight value generated randomly;
step 22): obtaining an optimal weight value through the particle swarm algorithm;
step 23): forming a neural network structure according to the obtained optimal weight;
step 24): carrying out neural network training and calculating an output error;
step 25): updating the weight according to a learning algorithm;
step 26): whether the condition of reaching the maximum iteration times is met or not is judged, if yes, the training is finished, and the optimized compensation control rate is output; otherwise, go back to step 24).
Compared with the prior art, the invention has the following obvious advantages:
1) the sliding mode controls the variable quantity corresponding to the added parameter, so that the robustness of the system when the parameter is changed is improved;
2) the method adopts the Elman neural network to estimate the value of the uncertainty factor in the system, replaces the switching control in the traditional sliding mode control, can weaken the buffeting phenomenon of the sliding mode control, further improves the robustness of the system and improves the power factor of the circuit;
3) the invention adopts the particle swarm optimization algorithm to realize the online optimization of the weight of the Elman neural network, and overcomes the defects of low convergence speed and easy falling into local extreme values during the neural network training.
Drawings
Fig. 1 is a schematic diagram of a main circuit topology and a current path of a Boost APFC;
FIG. 2 is a diagram of the Elman neural network architecture;
FIG. 3 is a flow chart of a particle swarm optimization algorithm;
FIG. 4 is a flow chart of PSO optimizing Elman neural network weights;
fig. 5 is a block diagram of a Boost APFC control system.
Detailed Description
The invention is further described below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
As shown in fig. 1, the Boost APFC main circuit topology and current path:
the main circuit adopts a Boost APFC topological structure, consists of a single-phase bridge type uncontrolled rectifier and a DC-DC Boost converter and comprises an input capacitor C1Boost inductor L, switch tube T, diode D and output capacitor CoAnd a load R. v. ofin＝Vs| sin ω t | is the output voltage of the uncontrolled rectifier bridge, VsIs the peak value of the grid voltage; r isLIs an inductance equivalent resistance, iLIs an inductive current, voOutputting voltage for direct current;
when the switch tube T is conducted, the inductor L stores energy and the capacitor CoFor supplying power to the load R, the current path is as shown by dotted lines ① and ② in fig. 1, and when the equivalent series resistance of the capacitor is neglected, according to KVL and KCL, the following results are obtained:
when the switch tube T is turned off, the inductor L is simultaneously the capacitor CoAnd load R, the current path is shown as dashed line ③ in fig. 1, and is available from KVL and KCL:
the state of the switch T is indicated by u, and when the switch T is turned on, u is 1; when T is off, u is 0. Equations (1) and (2) can be simplified as:
for the current expression in equation (3), L, r is idealL、vinAll are constant values, but considering the unknown parameter variation in practical application, it is expressed as a nominal value plus a variation:
for ease of analysis, the above equation can be simplified to:
wherein L isn、rLnIs a nominal value, Δ L, Δ rLn、ΔvinFor the variation, ρ is an unknown item in the system, and includes uncertainty of system parameter disturbance. Assuming that the boundary of ρ is given, | ρ<γ, γ is the positive boundary value of uncertainty vector ρ.
A sliding mode controller:
the purpose of PFC current sliding mode control is to enable an inductive current iLTracking reference current irefTherefore, the tracking error of the inductor current is defined as ei＝iref-iLThen eiThe derivative of (d) can be expressed as:
wherein ev＝vref-voIndicating output voltage tracking error, vrefIn order to output a voltage reference to the voltage,
the design integral type sliding surface is as follows:
wherein, z (e)i) Is required to satisfyλ is a non-zero positive constant. To achieve the desired sliding mode control of the system, i.e. to achieve the desired sliding mode controlAnd in the case of no external interference (rho is 0), if the value is equal toIf the solution of u exists, the solution is called the equivalent control u of the system in the sliding mode areaeq。
Order to(ρ ═ 0), by the formula (8):
for a system with uncertainty and external interference, the control rate is generally equal to the control ueqBased on the above, a switching control u is addedvssTo achieve robust control of the system. The system control rate uconThe design is as follows:
wherein,sgn(s) represents a sign function. Substituting the formula (10) into the formula (6) can obtain
The starting of the sliding mode movement is premised on the existence and the accessibility of a sliding mode, and the essential condition for the existence of the sliding mode isAnd isThe equivalent of the arrival condition isThis is expressed by the lyapunov function, and the arrival conditions can be rewritten as:and is
According to (8) and (11), the derivative of V is:
due to the fact thatS | ≧ 0, satisfy γ>In the case of | ρ | there is,namely, it isIs a negative definite function. According to Lyapunov's theorem, the tracking error e of the inductive currentiThe gradual trend is zero, and the gradual stability of the sliding mode control system and the sliding mode motion in the whole control period can be ensured whether system uncertainty exists or not.
Elman neural network:
the Elman neural network is adopted to estimate the value of the uncertainty factor in the system, switching control is replaced, the buffeting phenomenon of sliding mode control can be weakened, and the system robustness is improved.
Let the total control rate be ucon＝ueq+unnWhereinFor compensation control rate based on Elman neural network, yoIs the output of the Elman neural network and is used for realizing compensation control on uncertainty and external interference.
The Elman neural network structure is shown in fig. 2 and is composed of an input layer (i), a hidden layer (h), a carrying layer (c) and an output layer (o). The input to the neural network is xi(k)＝layeri(i＝1,2)，xi(k) And layeriRespectively representing the output and input of input level node i at time k, where x1＝ei，The output of the current moment obtained by the linear weighting of the input layer and the output of the previous moment of the hidden layer obtained by the feedback of the receiving layer jointly form the input of the hidden layer, which can be expressed as:
wherein x isc(k) Is the output of the receiving layer; omegaihIs the weight vector from the input layer to the hidden layer; n is the number of neurons (nodes) of the hidden layer (i.e., the carry over layer).
The Sigmoid function is denoted by g (·), then the output of the implicit layer of the function is:
the receiving layer is used for storing the state of the hidden layer at the previous moment, the capability of the network for processing dynamic information can be improved, and the output of the receiving layer is represented as:
xc(k)＝xh(k-1) (15)
wherein x ishAnd (k-1) is the output of the hidden layer at the moment k-1, and is converted into the input of the receiving layer at the moment k after being delayed.
The output of the neural network is:
wherein, yo(k) Is the output of an output layer neuron; layeroIs an input to the output layer; omegahoIs the weight vector from the hidden layer to the output layer.
The error function is defined as:
the learning rule of the weight of the neural network is as follows:
ω(k+1)＝ω(k)+Δω(k) (18)
ωihand ωhoThe correction amounts of (a) are:
wherein, ηih、ηhoAre respectively omegaihAnd ωhoThe learning rate of (a) is determined,in the training process, the error gradient of the weight is determined by utilizing the reverse transmission of the error, and the correction of the weight is realized.
According to the output error of the Elman neural network, defining a fitness function as:
wherein,and yoRepresenting the desired output and the actual output of the neural network, respectively; and m is the number of training samples of the neural network. The smaller the error is, the larger the fitness function value is, and the better the fitness is.
Particle swarm optimization algorithm:
when the dynamic learning rule is adopted to train the weight of the neural network, the convergence speed is slow, and the neural network is easy to fall into a local extremum. In order to improve the deficiency, a Particle Swarm Optimization (PSO) algorithm is adopted to realize the online optimization of the structural parameters of the Elman neural network.
Suppose that there are M particles in an N-dimensional search space, where the particles fly at a certain velocityThe jth particle is represented by Xj＝(Xj1,Xj2,…,XjN) At a flying speed of Vj＝(Vj1,Vj2,…,VjN) The optimal position (i.e. the individual extremum) through which the particle flies is Pj＝(Pj1,Pj2,…,PjN) The optimal position (i.e. global extremum) among the positions where all particles pass through in the whole population is Pglobal＝(Pg1,Pg2,…,PgN). And (3) carrying out iterative update on the particle speed and the particle position:
Vjd(t+1)＝wVjd(t)+c1r1(Pjd(t)-Xjd(t))+c2r2(Pgd(t)-Xjd(t)) (22)
Xjd(t+1)＝Xjd(t)+Vjd(t+1) (23)
wherein w is the inertial weight; c. C1、c2The step length of the particle flying to the self optimal position and the global optimal position is respectively represented by a learning factor (acceleration constant); r is1、r2Is [0,1 ]]A random number generated in between, representing an uncertain disturbance in flight; t is the number of iterations; vjd(t)、Xjd(t) respectively representing the speed and position of the jth particle in the d-dimension at the t-th iteration; pjd(t) represents the position of the individual extremum of the d-th dimension of the jth particle at the tth iteration; pgd(t) represents the position of the global extremum of the population of particles in the d-th dimension at the t-th iteration.
w may be determined by linearly decreasing the inertia weight:
wherein, wmax、wminRespectively representing the maximum value and the minimum value of w, namely the initial value of the inertia weight and the value when the maximum iteration number is reached; t, tmaxRespectively represent the currentThe number of iterations and the maximum number of iterations.
The flight speed of the particles is limited to [ -v ]max,vmax]Within, when updated, the obtained velocity VjAnd when the speed is larger than or smaller than the limit value, the speed at the moment is equal to the upper limit value and the lower limit value.
The particle swarm optimization algorithm is shown in FIG. 3, and comprises the following steps:
step 1: initializing the population (generating an initial population of particles having a certain number of individuals, randomly initiating the velocity V of each individual particlejAnd position Xj)；
Step 2: calculating the individual fitness of the particles;
and step 3: the individual fitness of each particle is compared with the optimal passing position Pj(individual extremum) fitness is compared, if better than PjThe fitness of (2) then updates Pj(ii) a Otherwise, the original P is keptj；
And 4, step 4: the individual fitness of each particle and the optimal position P passed by the groupglobal(global extreme) fitness is compared, if better than PglobalThe fitness of (2) then updates Pglobal(ii) a Otherwise, the original P is keptglobal；
And 5: updating the inertial weight according to equation (24);
step 6: updating the velocity and position of the particle according to equations (22) and (23);
and 7: if the optimization condition is met or the maximum iteration number is reached, outputting; otherwise, repeating the steps 2-6 until the conditions are met.
As shown in fig. 4, the particle swarm optimization optimizes the implementation steps of the Elman neural network:
step 1: initializing the structure of Elman neural network and setting the parameter omegaih、ωhoEncoding as parameters of the particles, randomly generating an initial population of particles in a solution space, each encodingThe code string represents weight distribution, and the scale, position and speed of the particle swarm are initialized;
step 2: carrying out sample training on the neural network;
and step 3: calculating the fitness function value of each particle according to the formula (21) so as to determine the individual extreme value PjAnd a global extremum Pglobal；
And 4, step 4: respectively matching the individual fitness of the particles with an individual extreme value PjAnd a global extremum PglobalThe corresponding fitness is compared, if the fitness is better than the fitness, P is updatedjAnd Pglobal；
And 5: updating the inertial weight and the velocity and position of the particle according to the formulas (22) to (24);
step 6: judging whether the optimization target is met or the maximum iteration number is reached, and if the optimization target meets the conditions, performing a step 7; otherwise, obtaining a group of new particle swarms to generate a new Elman neural network, and returning to the step 2.
And 7: decoding the global extremum of the particle group to serve as an optimized Elman neural network parameter, and learning by the neural network;
and 8: the algorithm ends.
As shown in fig. 5, the overall control process of the system is as follows:
direct current output voltage vo and reference voltage vrefError e ofvPerforming proportional integral control to make vo track vrefOutput quantity m of PI control module and unit sine half-wave (| v) of grid voltages|/vs(pk)，|vsL is positive half-wave of network voltage, vs(pk)For grid voltage peaks) to generate a reference current irefWith the collected inductor current iLError e ofiA slip form surface s for slip form control is formed. Method for estimating compensation control rate u on line by adopting Elman neural networknnOptimizing weight parameters of the Elman neural network by adopting a particle swarm optimization algorithm to obtain a compensation control rate unnAdding the equivalent control rate obtained by sliding mode control to obtain the total sliding mode control rate uconAnd a triangular carrier signal vΔAfter comparison, a PWM signal is generated to drive the power switch tube T to be switched on and off.
The technical means disclosed in the invention scheme are not limited to the technical means disclosed in the above embodiments, but also include the technical scheme formed by any combination of the above technical features. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and such improvements and modifications are also considered to be within the scope of the present invention.
Claims (5)
1. The utility model provides a APFC control system of online compensation control rate, includes Boost APFC main circuit, sliding mode control circuit, and Boost APFC main circuit converts alternating current power supply rectification filter back into high voltage direct current, and sliding mode control circuit controls the inductive current and the output voltage who gathers from Boost APFC main circuit, realizes the power factor correction to obtain smooth stable output voltage, its characterized in that: the sliding mode control circuit comprises a voltage outer ring PI control module, a sliding mode controller, a PWM comparator and a driving circuit module; the sliding mode controller comprises a sliding mode surface module, a sliding mode control module, an Elman neural network algorithm module and a particle swarm algorithm module;
the voltage outer ring PI control module is used for generating reference current and transmitting the reference current to the sliding mode surface module;
the sliding mode surface module is used for forming a sliding mode surface of the sliding mode controller;
the Elman neural network algorithm module is used for estimating the compensation control rate on line, and the particle swarm algorithm module is adopted to optimize the weight of the Elman neural network to obtain the optimized compensation control rate;
the sliding mode control module is used for receiving the sliding mode surface obtained by the sliding mode surface module, generating an equivalent control rate, adding the equivalent control rate and the optimized compensation control rate obtained by the Elman neural network algorithm module, and outputting a total sliding mode control rate; wherein, the Elman neural network estimates the value of uncertainty factor in the control system, and the total sliding mode control rate is set asu con ＝u eq +u nn Whereinu nn For an optimized compensatory control rate based on the Elman neural network,u nn =C n -1 y o ，u eq in order to achieve an equivalent control rate,y o is the output of the Elman neural network and is used for realizing compensation control on uncertainty and applied interference, whereinC n =-v o /L n ，v o Is the dc output voltage of the main circuit of the Boost APFC,L n inductance of main circuit for Boost APFCLA nominal value of (d);
the PWM comparator is used for comparing the total sliding mode control rate with a preset triangular carrier signal to generate a PWM signal and transmitting the PWM signal to the driving circuit module;
the driving circuit module is used for driving the power switch tube of the Boost APFC main circuit to be switched on and off.
2. The APFC control system for compensating a control rate on line according to claim 1, wherein: the PI control module comprises a PI controller and a multiplier, the PI controller is used for carrying out proportional integral control on the error between the direct current output voltage and the reference voltage so that the direct current output voltage tracks the reference voltage, and the multiplier is used for receiving the output quantity of the PI controller and multiplying the output quantity by the unit sine half wave of the power grid voltage to generate reference current.
3. The APFC control system for compensating a control rate on line according to claim 1 or 2, wherein: the sliding mode surface module is used for calculating the error between the reference current and the inductive current of the Boost APFC main circuit to form a sliding mode surface of the sliding mode controller.
4. The APFC control system for compensating a control rate on line according to claim 1, wherein: the particle swarm optimization module for optimizing the weight of the Elman neural network comprises the following steps:
step 1): initializing the structure of an Elman neural network, encoding the weight vectors from a parameter input layer to a hidden layer and the weight vectors from the hidden layer to an output layer as parameters of particles, randomly generating an initial particle population in a solution space, wherein each encoding string represents weight distribution, and initializing the scale, position and speed of a particle swarm;
step 2): carrying out sample training on the neural network;
step 3): calculating a fitness function value of each particle, thereby determining an individual extremum and a global extremum;
step 4): comparing the individual fitness of the particles with the fitness corresponding to the individual extreme value and the global extreme value respectively, and updating the individual extreme value and the global extreme value if the individual fitness is better than the global extreme value;
step 5): updating the inertia weight by adopting a mode of linearly decreasing the inertia weight, and performing iterative update on the speed and the position of the particle;
step 6): judging whether the optimization target is met or the maximum iteration number is reached, and if the optimization target meets the conditions, performing step 7); otherwise, obtaining a group of new particle swarms to generate a new Elman neural network, and returning to the step 2);
step 7): and decoding the global extremum of the particle group to serve as an optimized Elman neural network parameter, learning by the neural network, and finishing optimization.
5. The APFC control system for compensating the control rate online according to claim 1 or 4, wherein: the method for obtaining the optimized compensation control rate by the Elman neural network algorithm module comprises the following steps:
step 21): acquiring an initial weight value generated randomly;
step 22): obtaining an optimal weight value through the particle swarm algorithm;
step 23): forming a neural network structure according to the obtained optimal weight;
step 24): carrying out neural network training and calculating an output error;
step 25): updating the weight according to a learning algorithm;
step 26): whether the condition of reaching the maximum iteration times is met or not is judged, if yes, the training is finished, and the optimized compensation control rate is output; otherwise, go back to step 24).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811167864.7A CN109217651B (en) | 2018-10-08 | 2018-10-08 | A kind of APFC control system of online compensation control rate |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811167864.7A CN109217651B (en) | 2018-10-08 | 2018-10-08 | A kind of APFC control system of online compensation control rate |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109217651A CN109217651A (en) | 2019-01-15 |
| CN109217651B true CN109217651B (en) | 2019-06-28 |
Family
ID=64983113
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811167864.7A Active CN109217651B (en) | 2018-10-08 | 2018-10-08 | A kind of APFC control system of online compensation control rate |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109217651B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110712539A (en) * | 2019-10-12 | 2020-01-21 | 南通大学 | Sliding mode control method for electric drive reconstruction type charger for electric automobile |
| CN111478592A (en) * | 2020-05-09 | 2020-07-31 | 哈尔滨理工大学 | Sliding mode control method of double-active full-bridge DC-DC converter |
| CN114362522A (en) * | 2022-01-12 | 2022-04-15 | 河南科技大学 | DC-DC converter control method applying fixed time control theory and particle swarm optimization |
| CN116760277A (en) * | 2023-06-05 | 2023-09-15 | 广州菲利斯太阳能科技有限公司 | Boost PFC circuit control method for photovoltaic inverter |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1086615A (en) * | 1992-11-02 | 1994-05-11 | 永大机电工业股份有限公司 | Be applicable to the module parallel power factor correcting method and the device of arbitrary load |
| CN103164742A (en) * | 2013-04-02 | 2013-06-19 | 南京邮电大学 | Server performance prediction method based on particle swarm optimization nerve network |
| CN108512411A (en) * | 2018-04-13 | 2018-09-07 | 贵州师范大学 | Digital large power totem PFC based on dual-integration sliding formwork control |
-
2018
- 2018-10-08 CN CN201811167864.7A patent/CN109217651B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1086615A (en) * | 1992-11-02 | 1994-05-11 | 永大机电工业股份有限公司 | Be applicable to the module parallel power factor correcting method and the device of arbitrary load |
| CN103164742A (en) * | 2013-04-02 | 2013-06-19 | 南京邮电大学 | Server performance prediction method based on particle swarm optimization nerve network |
| CN108512411A (en) * | 2018-04-13 | 2018-09-07 | 贵州师范大学 | Digital large power totem PFC based on dual-integration sliding formwork control |
Non-Patent Citations (2)
| Title |
|---|
| 基于Elman神经网络的永磁直线同步电机互补滑模控制;赵希梅等;《电工技术学报》;20180301;第33卷(第5期);第973-979页 |
| 开关电源PFC校正数字控制算法研究;王丹;《中国优秀硕士学位论文全文数据库 信息科技辑》;20171115(第11期);第43-54页 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109217651A (en) | 2019-01-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109217651B (en) | A kind of APFC control system of online compensation control rate | |
| CN109149923B (en) | APFC control system for estimating control rate on line | |
| Hajihosseini et al. | DC/DC power converter control-based deep machine learning techniques: Real-time implementation | |
| CN110806759B (en) | Aircraft route tracking method based on deep reinforcement learning | |
| Cabrera et al. | Tuning the stator resistance of induction motors using artificial neural network | |
| Lin et al. | Fuzzy neural adaptive tracking control of unknown chaotic systems with input saturation | |
| CN110048606B (en) | DC-DC boost converter dynamic sliding mode voltage control method based on interval two-type self-adaptive fuzzy neural network | |
| Siano et al. | Designing fuzzy logic controllers for DC–DC converters using multi-objective particle swarm optimization | |
| Norgaard et al. | Neural network prediction of new aircraft design coefficients | |
| CN105262358A (en) | Bidirectional grid connected inverter | |
| CN116914751B (en) | Intelligent power distribution control system | |
| Fan et al. | Multi-agent deep reinforcement learning-based distributed optimal generation control of DC microgrids | |
| Wu et al. | Adaptive neural predictive control for permanent magnet synchronous motor systems with long delay time | |
| Darvill et al. | An ANFIS-PI based boost converter control scheme | |
| Merrikh-Bayat et al. | Comparing the performance of optimal PID and optimal fractional-order PID controllers applied to the nonlinear boost converter | |
| Ufnalski et al. | A plug-in direct particle swarm repetitive controller for a single-phase inverter | |
| CN105867116A (en) | Electricity network harmonic current signal tracking control method based on time delay compensation | |
| Abiyev | Fuzzy wavelet neural network for prediction of electricity consumption | |
| Patil et al. | Analysis of power loss in forward converter transformer using a novel machine learning-based optimization framework | |
| Benmabrouk et al. | Speed control of DC machine using adaptive neural IMC controller based on recurrent neural network | |
| Frijet et al. | Adaptive neural network internal model control for PMSM speed regulation | |
| Colot et al. | Enhanced neural network-based polytopic model for large-signal black-box modeling of power electronic converters | |
| CN113629780B (en) | Microgrid power converter control method, system, storage medium and device | |
| CN105846431A (en) | Power harmonics current signal tracking control method | |
| CN115987086A (en) | Single-switch DC-DC converter on-line control method based on neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230531 Address after: 215500 Building 1, No.8, research institute road, Changshu Economic and Technological Development Zone, Suzhou City, Jiangsu Province Patentee after: Jiangsu Qixun Digital Information Technology Co.,Ltd. Address before: 223005 Huaian 1 Jiangsu economic and Technological Development Zone Patentee before: HUAIYIN INSTITUTE OF TECHNOLOGY |
|
| TR01 | Transfer of patent right |