CN116470576A - Multilevel inverter system based on PWM and fuzzy neural MPPT control - Google Patents
Multilevel inverter system based on PWM and fuzzy neural MPPT control Download PDFInfo
- Publication number
- CN116470576A CN116470576A CN202310444917.XA CN202310444917A CN116470576A CN 116470576 A CN116470576 A CN 116470576A CN 202310444917 A CN202310444917 A CN 202310444917A CN 116470576 A CN116470576 A CN 116470576A
- Authority
- CN
- China
- Prior art keywords
- voltage
- bridge arm
- current
- switch tube
- phase
- 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
- 230000001537 neural effect Effects 0.000 title claims abstract description 20
- 239000013598 vector Substances 0.000 claims abstract description 57
- 230000007935 neutral effect Effects 0.000 claims abstract description 25
- 239000003990 capacitor Substances 0.000 claims description 70
- 238000012549 training Methods 0.000 claims description 35
- 238000013528 artificial neural network Methods 0.000 claims description 22
- 238000003062 neural network model Methods 0.000 claims description 19
- 238000009826 distribution Methods 0.000 claims description 18
- 238000000034 method Methods 0.000 claims description 17
- 238000006243 chemical reaction Methods 0.000 claims description 15
- 230000001276 controlling effect Effects 0.000 claims description 10
- 238000005457 optimization Methods 0.000 claims description 8
- XOFYZVNMUHMLCC-ZPOLXVRWSA-N prednisone Chemical compound O=C1C=C[C@]2(C)[C@H]3C(=O)C[C@](C)([C@@](CC4)(O)C(=O)CO)[C@@H]4[C@@H]3CCC2=C1 XOFYZVNMUHMLCC-ZPOLXVRWSA-N 0.000 claims description 8
- 230000001105 regulatory effect Effects 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 6
- 230000008447 perception Effects 0.000 claims description 6
- 230000000295 complement effect Effects 0.000 claims description 5
- 230000007547 defect Effects 0.000 abstract description 4
- 230000006870 function Effects 0.000 description 12
- 238000010586 diagram Methods 0.000 description 11
- 210000002569 neuron Anatomy 0.000 description 7
- 230000009471 action Effects 0.000 description 6
- 210000004027 cell Anatomy 0.000 description 6
- 230000000946 synaptic effect Effects 0.000 description 6
- 230000004913 activation Effects 0.000 description 5
- 238000004146 energy storage Methods 0.000 description 4
- 230000006872 improvement Effects 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 210000005036 nerve Anatomy 0.000 description 3
- 210000004205 output neuron Anatomy 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000005855 radiation Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000010248 power generation Methods 0.000 description 2
- 238000011946 reduction process Methods 0.000 description 2
- 230000004584 weight gain Effects 0.000 description 2
- 235000019786 weight gain Nutrition 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 239000005431 greenhouse gas Substances 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 230000002277 temperature effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
- H02J3/381—Dispersed generators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/043—Architecture, e.g. interconnection topology based on fuzzy logic, fuzzy membership or fuzzy inference, e.g. adaptive neuro-fuzzy inference systems [ANFIS]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/088—Non-supervised learning, e.g. competitive learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/09—Supervised learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/04—Inference or reasoning models
- G06N5/048—Fuzzy inferencing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/02—Computing arrangements based on specific mathematical models using fuzzy logic
-
- 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
- H02M3/00—Conversion of dc power input into dc power output
- H02M3/02—Conversion of dc power input into dc power output without intermediate conversion into ac
- H02M3/04—Conversion of dc power input into dc power output without intermediate conversion into ac by static converters
- H02M3/10—Conversion of dc power input into dc power output without intermediate conversion into ac by static converters using discharge tubes with control electrode or semiconductor devices with control electrode
- H02M3/145—Conversion of dc power input into dc power output without intermediate conversion into ac by static converters using discharge tubes with control electrode or semiconductor devices with control electrode using devices of a triode or transistor type requiring continuous application of a control signal
- H02M3/155—Conversion of dc power input into dc power output without intermediate conversion into ac by static converters using discharge tubes with control electrode or semiconductor devices with control electrode using devices of a triode or transistor type requiring continuous application of a control signal using semiconductor devices only
-
- 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
- H02M7/00—Conversion of ac power input into dc power output; Conversion of dc power input into ac power output
- H02M7/42—Conversion of dc power input into ac power output without possibility of reversal
- H02M7/44—Conversion of dc power input into ac power output without possibility of reversal by static converters
- H02M7/48—Conversion of dc power input into ac power output without possibility of reversal by static converters using discharge tubes with control electrode or semiconductor devices with control electrode
- H02M7/483—Converters with outputs that each can have more than two voltages levels
-
- 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
- H02M7/00—Conversion of ac power input into dc power output; Conversion of dc power input into ac power output
- H02M7/42—Conversion of dc power input into ac power output without possibility of reversal
- H02M7/44—Conversion of dc power input into ac power output without possibility of reversal by static converters
- H02M7/48—Conversion of dc power input into ac power output without possibility of reversal by static converters using discharge tubes with control electrode or semiconductor devices with control electrode
- H02M7/53—Conversion of dc power input into ac power output without possibility of reversal by static converters using discharge tubes with control electrode or semiconductor devices with control electrode using devices of a triode or transistor type requiring continuous application of a control signal
- H02M7/537—Conversion of dc power input into ac power output without possibility of reversal by static converters using discharge tubes with control electrode or semiconductor devices with control electrode using devices of a triode or transistor type requiring continuous application of a control signal using semiconductor devices only, e.g. single switched pulse inverters
- H02M7/5387—Conversion of dc power input into ac power output without possibility of reversal by static converters using discharge tubes with control electrode or semiconductor devices with control electrode using devices of a triode or transistor type requiring continuous application of a control signal using semiconductor devices only, e.g. single switched pulse inverters in a bridge configuration
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
- H02J2300/20—The dispersed energy generation being of renewable origin
- H02J2300/22—The renewable source being solar energy
- H02J2300/24—The renewable source being solar energy of photovoltaic origin
- H02J2300/26—The renewable source being solar energy of photovoltaic origin involving maximum power point tracking control for photovoltaic sources
-
- 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
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/50—Photovoltaic [PV] energy
- Y02E10/56—Power conversion systems, e.g. maximum power point trackers
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Power Engineering (AREA)
- Automation & Control Theory (AREA)
- Fuzzy Systems (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Algebra (AREA)
- Inverter Devices (AREA)
Abstract
The invention discloses a multi-level inversion system based on PWM and fuzzy neural MPPT control, which is characterized by comprising: the control system, the circuit breaker, the load, the photovoltaic array, the booster circuit, the three-phase three-level inverter circuit and the filter circuit are sequentially connected; the circuit breaker is also connected with a power grid, and a public end generated by the connection of the circuit breaker and the power grid is connected with a control system; the control system outputs a first path to be connected with the circuit breaker, the control system outputs a second path to be connected with a first controller based on the modulated three-level to two-level space vector, and the first controller is connected with the PWM controller; the photovoltaic array is also connected with an MPPT controller. The embodiment of the invention can solve the defects of unbalanced voltage caused by unbalanced voltage and neutral voltage and current and neutral potential in the existing multi-level inverter, increased waveform distortion of output current, large conduction loss of an inversion IGBT, large ripple current around MPPT, slow MPPT tracking and low inversion efficiency.
Description
Technical Field
The invention relates to the technical field of electric power, in particular to a multi-level inversion system based on PWM and fuzzy neural MPPT control.
Background
The traditional energy sources are limited in number on earth and lead to massive degradation of natural resources. This deterioration results in pollution and greenhouse gas emissions. Photovoltaic has a long service life, minimal maintenance, and no noise or interference effects. However, photovoltaic is affected by solar energy generated by a photovoltaic cell panel, and the tracking and prediction of the maximum power generated by MPPT are affected by irradiation, sunlight incidence angle, cloud cover, wind, rain and snow, battery temperature and load conditions. The existing three-level inverter device has the defects that the voltage withstand voltage is only direct current bus voltage, the voltage deviation of a general DC capacitor is large, the waveform distortion of an inversion output current is increased due to unbalanced midpoint potential, the low-order harmonic quantity is increased, the voltage balance is difficult to realize, the IGBT conduction loss of the PWM-controlled inverter device is large, the main bridge arm works for a long time, the heat distribution temperature of a power device is high, the ripple current around MPPT is large, the MPPT operates slowly or the tracking efficiency is low.
Disclosure of Invention
The invention provides a multilevel inversion system based on PWM and fuzzy nerve MPPT control, which extracts maximum power from a photovoltaic array circuit and controls direct current bus voltage through the MPPT based on the fuzzy nerve, injects sine input current with unit power factor into a power grid based on integration proportional controller (PI) fusion control of modulating three-level to two-level space vector and PWM, balances voltage unbalance caused by neutral potential unbalance based on the fuzzy modulating three-level to two-level space vector, and solves the defects that the prior multilevel inverter does not comprehensively track and monitor weather, photovoltaic and neutral voltage and current, and voltage unbalance caused by neutral potential unbalance, the waveform distortion of output current is increased, the conduction loss of an inversion IGBT is large, the temperature rise is high, ripple current around the MPPT is large, and the MPPT tracking is slow and the inversion efficiency is low. Under the conditions of various environments such as unequal solar radiation of the photovoltaic cell strings, the sine curve of the output current of the multi-level inverter is kept, the dynamic performance of the multi-level inverter is improved under different atmospheric conditions, the inversion efficiency is improved, and the THD of the electric energy quality is reduced.
To achieve the above object, an embodiment of the present invention provides a multi-level inverter system based on PWM and fuzzy neural MPPT control, including: the control system, the circuit breaker, the load, the photovoltaic array, the booster circuit, the three-phase three-level inverter circuit and the filter circuit are sequentially connected; wherein,,
the photovoltaic array includes: a first photovoltaic array and a second photovoltaic array;
the booster circuit includes: a first booster circuit and a second booster circuit;
the first booster circuit includes: capacitor C PV1 Inductance L 1 Switch tube S 1 And diode D 1 ；
The second booster circuit includes: capacitor C PV2 Inductance L 2 Switch tube S 2 And diode D 2 ；
The three-phase three-level inverter circuit comprises: capacitor C in series 1 And capacitor C 2 A capacitor C connected in series with 1 And capacitor C 2 The device comprises an A-phase inverter circuit, a B-phase inverter circuit and a C-phase inverter circuit which are connected in parallel;
the filter circuit includes: the first filter circuit, the second filter circuit and the third filter circuit are connected in parallel;
first photovoltaic array and capacitor C PV1 The first photovoltaic array outputs the positive electrode and the inductance L in parallel connection 1 Connected to, inductance L 1 The other end and the switching device S 1 Collector, diode D of (c) 1 Is connected with the anode of diode D 1 Cathode and capacitor C of (2) 1 One end of (C) is connected to 1 And the other end of (C) and the capacitor C 2 One end of (C) is connected to 2 And the other end of the switch tube S 2 Emitter connection of (c)
Second photovoltaic array and capacitor C PV2 The second photovoltaic array outputs the positive electrode and the inductance L in parallel connection 2 Connected to, inductance L 2 The other end and the switching device S 2 Collector, diode D of (c) 2 Is connected with the anode of diode D 2 Is of (2)Pole and switching device S 1 Emitter connection of (a);
capacitor C 1 And the other end of (C) and the capacitor C 2 Is connected to the common terminal and diode D 2 Cathode of (d) and switching device S 1 Is connected with a common terminal generated by the emitter connection;
the A-phase inverter circuit includes: four bridge arm switching tubes S connected in series in sequence a1 、S a2 、S a3 、S a4 And two clamping tubes D connected in series a1 、D a2 ；
The B-phase inverter circuit includes: four bridge arm switching tubes S connected in series in sequence b1 、S b2 、S b3 、S b4 And two clamping tubes D connected in series b1 、D b2 ；
The C-phase inverter circuit includes: four bridge arm switching tubes S connected in series in sequence c1 、S c2 、S c3 、S c4 And two clamping tubes D connected in series c1 、D c2 ；
Bridge arm switch tube S a1 And bridge arm switch tube S a2 Common terminal and bridge arm switch tube S a3 And bridge arm switch tube S a4 Two clamping tubes D connected in series are connected in parallel between the common ends of the two clamping tubes a1 、D a2 Clamping tube D a1 And clamping tube D a2 Is connected to the common terminal of the capacitor C 1 And capacitor C 2 Is connected with the public end of the box body;
bridge arm switch tube S b1 And bridge arm switch tube S b2 Common terminal and bridge arm switch tube S b3 And bridge arm switch tube S b4 Two clamping tubes D connected in series are connected in parallel between the common ends of the two clamping tubes b1 、D b2 Clamping tube D b1 And clamping tube D b2 Is connected to the common terminal of the capacitor C 1 And capacitor C 2 Is connected with the public end of the box body;
bridge arm switch tube S c1 And bridge arm switch tube S c2 Common terminal and bridge arm switch tube S c3 And bridge arm switch tube S c4 Two clamping tubes D connected in series are connected in parallel between the common ends of the two clamping tubes c1 、D c2 Clamping tube D c1 And clamping tube D c2 Is connected to the common terminal of the capacitor C 1 And capacitor C 2 Is connected with the public end of the box body;
bridge arm switch tube S a2 And bridge arm switch tube S a3 The public end generated by connection is output to a first filter circuit;
bridge arm switch tube S b2 And bridge arm switch tube S b3 The public end generated by connection is output to a second filter circuit;
bridge arm switch tube S c2 And bridge arm switch tube S c3 The public end generated by connection is output to a third filter circuit;
the first output end of the filter circuit is connected with the load, and the second output end of the filter circuit is connected with the circuit breaker;
the circuit breaker is also connected with a power grid, and a public end generated by the connection of the circuit breaker and the power grid is connected with the control system;
the output of the control system is connected with the circuit breaker in a first way, the output of the control system is connected with a first controller based on the modulated three-level to two-level space vector in a second way, and the first controller is connected with a PWM controller;
The first controller is used for adjusting midpoint potential by adopting a mixed midpoint potential balance algorithm based on fuzzy optimization, adopting a seven-segment switching sequence method for the area containing the middle vector, and adopting a fuzzy optimization time factor distribution method for the area without the middle vector;
the PWM controller is used for controlling on-off inversion of bridge arm switching tubes of the three-phase three-level inverter circuit;
the controller system is also connected with an MPPT controller;
the MPPT controller is used for obtaining a reference voltage V corresponding to the maximum power point based on the current state data and a preset neural network model REF And reference current F REF And based on reference voltage V REF And reference current I REF Obtaining a PWM control signal to track the maximum power; wherein the status data comprises: irradiance, air temperature, photovoltaic temperature, wind speed, open circuit voltage, short circuit current, photovoltaic input voltage, photovoltaicThe neural network model is obtained by training according to historical state data of the system under different working conditions and corresponding output voltage and current of the corresponding system running at a maximum power point.
As an improvement of the above scheme, the controlling the on-off inversion of the bridge arm switching tube of the three-phase three-level inverter circuit includes:
Bridge arm switch tube S with j phase j1 、S j2 、S j3 、S j4 The switch states of (a) are q respectively j1 、q j2 、q j3 、q j4 Then switch state q j1 、q j3 Complementary, switch state q j2 、q j4 Complementation;
capacitance C 2 The voltage V generated C2 And (V) C1 +V C2 ) The result of the comparison of/2 is input into a first PI controller which outputs a distribution ratio mu, the distribution ratio mu is input into the voltage V h Controller, voltage V h The controller obtains the input voltage V according to a preset equation h And will input voltage V h Respectively adding to three reference voltages;
by applying a DC port voltage V dc With a target voltage V dc * The result of comparison output is input into a second PI controller, and the second PI controller outputs an inverter modulation index m;
generating three reference voltages V based on an inverter modulation index m ao *、V bo *、V co *，V ao *、V bo * And V co * Regulating the voltage of a single capacitor, outputting a voltage V jo *；
Voltage V jo * And input voltage V h The three reference voltages added respectively are compared to generate a new voltage V jo_ * Will be voltage V jo_ * The positive electrode of the first comparator and the positive electrode of the second comparator are respectively input, the positive electrode of the first comparator is compared with the negative electrode of the single-phase circuit, and the switch state q is obtained by outputting j1 The positive pole of the second comparator is compared with the negative pole of the single-phase circuit, and the switch state q is obtained by output j2 ；
Based on the output, the obtained switch state q j1 、q j2 And controlling the on-off inversion of bridge arm switching tubes of the three-phase three-level inversion circuit.
As a modification of the above, the equation is:
V jo- R 1 i j -L j (di j /d t )-v on ＝0
v on ＝(v ao +v bo +v co )/3
v jo ＝(q j1 +q j2 -1)(v c1 +v c2 )/2
V h ＝(1-2μ)V DC /2+μV max +(1-μ)V min
wherein V is C2 For the voltage generated by the capacitor C2, V C1 Voltage q generated for capacitor C1 j1 Is a j-phase bridge arm switch tube S j1 Switch state, q j2 Is a j-phase bridge arm switch tube S j2 Switch state of V jo Voltage representing j phase, R 1 i j Represents the product of the first resistance (O passing through C) and the current of the corresponding phase j, L j (di j /d t ) Representing the product of the inductance of the respective phase and the current through the coil per unit time, v on The neutral line voltage of three phases, v ao 、v bo 、v co Represents the voltages of the phase output lines of the phase A, the phase B and the phase C respectively, mu is the distribution ratio, V dc Representing the dc port voltage.
As an improvement of the above scheme, the reference voltage V corresponding to the maximum power point is obtained based on the current state data and a preset neural network model REF And referenceCurrent I REF And based on reference voltage V REF And reference current I REF Obtaining a PWM control signal for maximum power tracking, including:
acquiring current state data;
obtaining a reference voltage V corresponding to the maximum power point based on the state data and a preset neural network model REF And reference current I REF And will reference voltage V REF And reference current I REF Switching to a second duty cycle;
according to the photovoltaic input power in the current state data, calculating an error E or an error variable delta E, inputting the error E or the error variable delta E into a fuzzy controller, outputting a duty cycle delta D by the fuzzy controller, converting the duty cycle delta D into an analog quantity through a fuzzification stage, outputting a first duty cycle after the analog quantity is converted by A/D, comparing the first duty cycle with a second duty cycle, inputting a comparison result into a grid driver, generating a PWM signal, driving DC/DC conversion by the PWM signal, driving MPPT by the DC/DC conversion, and driving the output of the maximum power of the photovoltaic array by the MPPT.
As an improvement of the above scheme, the reference voltage V corresponding to the maximum power point is obtained based on the state data and a preset neural network model REF And reference current I REF Comprising:
inputting the current irradiance to a preset irradiance prediction model to obtain predicted irradiance; the irradiance prediction model is obtained by training through a multi-layer perception layer device based on historical irradiance;
inputting current air temperature data into a preset air temperature prediction model to obtain predicted air temperature; the air temperature prediction model is obtained by training through an Eerman network based on historical air temperature;
Inputting current photovoltaic temperature data into a preset photovoltaic temperature prediction model to obtain predicted photovoltaic temperature; the photovoltaic temperature prediction model is obtained by training through an Elman network based on historical photovoltaic temperature;
inputting the current clouds into a preset clouds prediction model to obtain predicted clouds; the cloud prediction model is based on historical clouds and is obtained by training through a back propagation neural network;
inputting predicted irradiance, predicted air temperature, predicted photovoltaic temperature, predicted cloud, current wind speed, current open-circuit voltage, current short-circuit current, current photovoltaic input voltage, photovoltaic input current and current photovoltaic input power into the neural network model to obtain a reference voltage V corresponding to a maximum power point REF And reference current I REF 。
As an improvement of the above solution, the multi-level inversion system based on PWM and fuzzy neural MPPT control, the first filter circuit includes: resistor R A Inductance L A ；
The second filter circuit includes: resistor R B Inductance L B ；
The third filter circuit includes: resistor R C Inductance L C ；
Then, bridge arm switch tube S a2 And bridge arm switch tube S a3 The common terminal generated by connection is input to a resistor R A One end of the resistor R A And the other end of (2) is connected with inductance L A Is connected with one end of the connecting rod;
bridge arm switch tube S b2 And bridge arm switch tube S b3 The common terminal generated by connection is input to a resistor R B One end of the resistor R B And the other end of (2) is connected with inductance L B Is connected with one end of the connecting rod;
bridge arm switch tube S c2 And bridge arm switch tube S c3 The common terminal generated by connection is input to a resistor R C One end of the resistor R C And the other end of (2) is connected with inductance L C Is connected to one end of the connecting rod.
Compared with the prior art, the multi-level inversion system based on PWM and fuzzy neural MPPT control provided by the embodiment of the invention has the following beneficial effects:
(1) According to the embodiment of the invention, the maximum power is extracted from the photovoltaic array circuit and the direct current bus voltage is controlled through the MPPT based on the fuzzy nerve, the sine input current with the unit power factor is injected into the power grid based on the integration control of the integrated proportional controller (PI) for modulating the three-level to two-level space vector and PWM, and the voltage unbalance caused by the neutral point potential unbalance is balanced based on the fuzzy modulation three-level to two-level space vector, so that the defects that the current multi-level inverter does not comprehensively track and monitor weather, photovoltaic and neutral point voltage and current, and the voltage unbalance caused by the neutral point potential unbalance are solved, the waveform distortion of the output current is increased, the conduction loss of the inversion IGBT is large, the temperature rise is high, the ripple current around the MPPT is large, and the MPPT tracking is slow and the inversion efficiency is low. Under the conditions of various environments such as unequal solar radiation of the photovoltaic cell strings, the sine curve of the output current of the multi-level inverter is kept, the dynamic performance of the multi-level inverter is improved under different atmospheric conditions, the inversion efficiency is improved, and the THD of the electric energy quality is reduced.
(2) Compared with the prior art, the neural network algorithm adopted by the photovoltaic array power generation greatly reduces training samples, shortens the time required by the reduction process by months or years, eliminates mismatching of the training input and output of the prior neural network, timely interpolates or extrapolates a certain number of inputs and outputs to achieve matching, and scales time parameters arbitrarily, thereby effectively solving the problem that the maximum power tracking, short time, medium time and long-term energy storage duration of the photovoltaic are not adjustable. The fuzzy neural network greatly improves the tracking prediction accurate time length and accuracy, stability and reliability of the photovoltaic MPPT and the photovoltaic tracking conversion efficiency.
Drawings
Fig. 1 is a block diagram of a multi-level inverter system based on PWM and fuzzy neural MPPT control according to an embodiment of the present invention;
fig. 2 is a block diagram of a multi-level inverter system under PWM and fuzzy neural MPPT control according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of three-level to two-level space vector control based on adjustment according to an embodiment of the present invention;
fig. 4 is a schematic diagram of a switching vector when three-level to two-level space vectors n=1 and n=1 according to an embodiment of the present invention;
FIG. 5 is a block diagram of an inverter according to an embodiment of the present invention;
fig. 6 is a flow chart of an MPPT controller provided by an embodiment of the present invention;
fig. 7 is a schematic diagram of a topology structure of a multi-layer aware network module according to an embodiment of the present invention;
fig. 8 is a schematic diagram of a topology structure of an artificial neural module of a predictive gamma memory according to an embodiment of the invention.
Fig. 9 is a schematic diagram of an MPPT controller simulation inverter power structure according to an embodiment of the present invention;
fig. 10 is a schematic diagram of a topology structure of a three-phase three-level inverter and an inversion harmonic test structure of a control system according to an embodiment of the present invention;
FIG. 11 is a diagram showing waveforms of output voltage and current of a load according to an embodiment of the present invention;
fig. 12 is a waveform diagram of a load output three-level to two-level output voltage and current according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1 to 2, the multi-level inversion system based on PWM and fuzzy neural MPPT control includes: the control system 7, the circuit breaker 6, the load 5, the photovoltaic array 1, the booster circuit 2, the three-phase three-level inverter circuit 3 and the filter circuit 4 are sequentially connected; wherein,,
the photovoltaic array 1 includes: a first photovoltaic array and a second photovoltaic array;
the booster circuit 2 includes: a first booster circuit and a second booster circuit;
the first booster circuit includes: capacitor C PV1 Inductance L 1 Switch tube S 1 And diode D 1 ；
The second booster circuit includes: capacitor C PV2 Inductance L 2 Switch tube S 2 And diode D 2 ；
The three-phase three-level inverter circuit 3: capacitor C in series 1 And capacitor C 2 A capacitor C connected in series with 1 And capacitor C 2 The device comprises an A-phase inverter circuit, a B-phase inverter circuit and a C-phase inverter circuit which are connected in parallel;
the filter circuit 4 includes: the first filter circuit, the second filter circuit and the third filter circuit are connected in parallel;
first photovoltaic array and capacitor C PV1 The first photovoltaic array outputs the positive electrode and the inductance L in parallel connection 1 Connected to, inductance L 1 The other end and the switching device S 1 Collector, diode D of (c) 1 Is connected with the anode of diode D 1 Cathode and capacitor C of (2) 1 One end of (C) is connected to 1 And the other end of (C) and the capacitor C 2 One end of (C) is connected to 2 And the other end of the switch tube S 2 Emitter connection of (c)
Second photovoltaic array and capacitor C PV2 The second photovoltaic array outputs the positive electrode and the inductance L in parallel connection 2 Connected to, inductance L 2 The other end and the switching device S 2 Collector, diode D of (c) 2 Is connected with the anode of diode D 2 Cathode of (d) and switching device S 1 Emitter connection of (a);
capacitor C 1 And the other end of (C) and the capacitor C 2 Is connected to the common terminal and diode D 2 Cathode of (d) and switching device S 1 Is connected with a common terminal generated by the emitter connection;
the A-phase inverter circuit includes: four bridge arm switching tubes S connected in series in sequence a1 、S a2 、S a3 、S a4 And two clamping tubes D connected in series a1 、D a2 ；
The B-phase inverter circuit includes: four bridge arm switching tubes S connected in series in sequence b1 、S b2 、S b3 、S b4 (not shown) and two clamping tubes D connected in series b1 、D b2 (not shown);
the C-phase inverter circuit includes: four bridge arm switching tubes S connected in series in sequence c1 、S c2 、S c3 、S c4 (not shown) and two clamping tubes D connected in series c1 、D c2 (not shown);
bridge arm switch tube S a1 And bridge arm switch tube S a2 Common terminal and bridge arm switch tube S a3 And bridge arm switch tube S a4 Two clamping tubes D connected in series are connected in parallel between the common ends of the two clamping tubes a1 、D a2 Clamping tube D a1 And clamping tube D a2 Is connected to the common terminal of the capacitor C 1 And capacitor C 2 Is connected with the public end of the box body;
bridge arm switch tube S b1 And bridge arm switch tube S b2 Common terminal and bridge arm switch tube S b3 And bridge arm switch tube S b4 Two clamping tubes D connected in series are connected in parallel between the common ends of the two clamping tubes b1 、D b2 Clamping tube D b1 And clamping tube D b2 Is connected to the common terminal of the capacitor C 1 And capacitor C 2 Is connected with the public end of the box body;
bridge arm switch tube S c1 And bridge arm switch tube S c2 Common terminal and bridge arm switch tube S c3 And bridge arm switch tube S c4 Two clamping tubes D connected in series are connected in parallel between the common ends of the two clamping tubes c1 、D c2 Clamping tube D c1 And clamping tube D c2 Is connected to the common terminal of the capacitor C 1 And capacitor C 2 Is connected with the public end of the box body;
bridge arm switch tube S a2 And bridge arm switch tube S a3 The public end generated by connection is output to a first filter circuit;
bridge arm switch tube S b2 And bridge arm switch tube S b3 The public end generated by connection is output to a second filter circuit;
bridge arm switch tube S c2 And bridge arm switch tube S c3 The public end generated by connection is output to a third filter circuit;
a first output end of the filter circuit 4 is connected with the load 5, and a second output end of the filter circuit 4 is connected with the circuit breaker 6;
the circuit breaker 6 is also connected with a power grid, and a public end generated by the connection of the circuit breaker 6 and the power grid is connected with the control system;
the output of the control system 7 is connected with the circuit breaker 6 in a first way, the output of the control system 7 is connected with a first controller 8 which is based on the modulated three-level to two-level space vector, and the first controller 8 is connected with a PWM controller 9;
The first controller 8 is configured to adjust a midpoint potential by using a mixed midpoint potential balancing algorithm based on fuzzy optimization, and adjust a midpoint potential by using a seven-segment switching sequence method for a region containing a middle vector, and adjust a midpoint potential by using a fuzzy optimization time factor distribution method for a region not containing a middle vector;
the PWM controller 9 is used for controlling on-off inversion of bridge arm switching tubes of the three-phase three-level inverter circuit;
the controller system is also connected with an MPPT controller 10;
the MPPT controller 10 is configured to obtain a reference voltage V corresponding to a maximum power point based on the current state data and a preset neural network model REF And reference current I REF And based on reference voltage V REF And reference current I REF Obtaining a PWM control signal to track the maximum power; wherein the status data comprises: irradiance, air temperature, photovoltaic temperature, wind speed, open-circuit voltage, short-circuit current, photovoltaic input voltage, photovoltaic input current and photovoltaic input power, wherein the neural network model is obtained through training according to historical state data of a system under different working conditions and corresponding output voltage and current of the corresponding system running at a maximum power point.
Specifically, the first filter circuit includes: resistor R A Inductance L A ；
The second filter circuit includes: resistor R B Inductance L B ；
The third filter circuit includes: resistor R C Inductance L C ；
Then, bridge arm switch tube S a2 And bridge arm switch tube S a3 The common terminal generated by connection is input to a resistor R A One end of the resistor R A And the other end of (2) is connected with inductance L A Is connected with one end of the connecting rod;
bridge arm switch tube S b2 And bridge arm switch tube S b3 The common terminal generated by connection is input to a resistor R B One end of the resistor R B And the other end of (2) is connected with inductance L B Is connected with one end of the connecting rod;
bridge arm switch tube S c2 And bridge arm switch tube S c3 The common terminal generated by connection is input to a resistor R C One end of the resistor R C And the other end of (2) is connected with inductance L C Is connected to one end of the connecting rod.
Specifically, the PWM controller employs a proportional controller (PI) controller to generate SVPWM signal switches for IGBTs, thereby generating and regulating a 50Hz sinusoidal ac output voltage, and achieving high dynamic performance with low Total Harmonic Distortion (THD).
The three-phase three-level inverter circuit topology structure is that the neutral point clamping multi-level inverter A, B, C phases are respectively input by parallel photovoltaic arrays (C pv1 ，C pv2 ) Is boosted by the first path of photovoltaic (L) 1 ,S 1 ,D 1 ,,C 1 (ii) and a second path of photovoltaic boosting (L) 2 ,S 2 ,D 2 ,,C 2 (C) and a three-level neutral DC charging bus 1 ,C 2 ) Connection, three-level neutral DC charging bus (C) 1 ,C 2 ) Four bridge arm switches connected in series (S a1 ,S a2 ,S a3 ,S a4 And 2 bridge arm switches (D) a1 ，D a2 ) Neutral point clamping connection, 2 diode neutral points of first path photovoltaic boosting and second path photovoltaic boosting, and three-level neutral position direct current charging bus capacitor C 1 And C 2 Neutral point interconnection, arm switch (D) a1 ) Connection and bridge arm switch (S) a1 ,S a2 ) Neutral position, arm switch (D) a2 ) Connection and bridge arm switch (S) a3 ,S a4 ) The middle position is connected to form an A-phase inverter, and the B-phase inverter and the C-phase inverter have the same structure as the A-phase inverter to form a three-phase inverter circuit;
the phase B and phase C inverter circuits are the same as the phase A inverter circuits, and the bridge arm switch of the phase A inverter (S a3 ,S a4 ) Middle position output and L a, ，C a, The filter circuit is formed and outputs one path of output signal which is connected with a load, the other path of output signal is connected with a circuit breaker, and the circuit breaker is connected with a power grid; the circuit breaker is connected with a control system through a PCC at a common point in the middle of the power grid; the control system outputs a first path to be connected with the circuit breaker, the control system outputs a second path to be connected with a first controller based on the space vector from the three level to the two level, the first controller is connected with a PWM controller, and the PWM controller is connected with an output to control the on-off inversion of a bridge arm switch of a three-phase inverter circuit; and the current output by the inverter circuit is input into a control system through a first path of current after LC filtering, and the current voltage of the PCC at the common point of the power grid and the circuit breaker is input into the control system through a second path of current voltage.
Specifically, as shown in fig. 3, the modulation-based three-level to two-level space vector control includes dividing a three-level space vector large hexagon into 6 regions (I-V), dividing the three-level space vector large hexagon into six two-level small hexagons, overlapping the two-level small hexagons of n=1 and n=2 with a first region of the large hexagon, dividing the first region of the large hexagon (triangle region I) into regions a, B, C, D, and overlapping the center distances of the small hexagons and the large hexagons to be equal to V dc， The reference voltage of PCC is arranged in the first area and falls into a small hexagon with N=1 and two levels, and the offset vector U is subtracted 1， Obtaining a new space vector U ref 、U` 1 、U` 7 、U` 13 Then obtaining U' through space vector calculation of two levels 1 、U` 7 、U` 13 Time of action t of space vector 1 ，t 7 ，t 13 ；T s For sampling time, m is modulation index
U ref T s ＝U 1 T 1 +U 7 T 7 +U 13 T 13
T s ＝T 1 +T 7 +T 13
````
U ref T s ＝U 1 T 1 +U 7 T 7 +U 13 T 13
When 0 > m < 0.5 and 0.5 > m < 0.577 are positioned in the area A, the area A comprises zero vectors and small vectors, a pair of paired redundant small vectors exist in three levels, and the influence of the paired points is counteracted; the invention adopts a control factor f for regulating the acting time of positive and negative small vectors by a non-redundant paired small vector to perform neutral potential control, the total quantity of charges flowing through neutral point bit is equal to 0, the neutral vector in the area A is modulated by adopting a fuzzy control time factor to distribute, capacitance deviation e and the variation rate ec of deviation are used as input of a fuzzy controller, the balance factor is fuzzy output combination, the input and output meet +.f (n) = +.f (e, ec). The argument of quantized input variables for capacitance deviation and deviation variation rate is +/-1, and the scale factor of output balance factor +. ∫ g∫(n)∈[-1 1]，К` ∫ As a scale factor, when the capacitance voltage DeltaV dc When the vector is more than 0, selecting positive small vector acting time (1- < DEG > C/2), wherein the negative small vector acting time is (1 + < DEG >/2, and delta V dc When the value is less than 0, selecting positive small vector acting time (1 + [ factor)/2 and negative small vector acting time (1- [ factor)/2, defining that capacitance voltage deviation and deviation change rate are 5 fuzzy subsets (NB, NS, ZE, PS, PB), respectively representing negative big, negative small, zero, positive small, positive big and membership functions, selecting trapezoidal curves, adopting a mamdanni method for fuzzy reasoning, and adopting a gravity center method for defuzzification. The membership function and the output function are respectively:
the fuzzy inference rules are shown in table 1:
TABLE 1 fuzzy inference rules
The vectors are located in two-level space vectors n=1, and the switching schematic of the sector is shown in fig. 4, and the implementation of mixed midpoint potential based on fuzzy optimization. Firstly, judging whether a middle vector is contained or not (the middle vector is not contained in the area A, the middle vector is contained in other areas), when the middle vector is contained in the area A, adopting seven-section switching sequences to adjust the balance of midpoint potential, judging which of two adjacent overlapped areas is used for adjustment according to the difference value of capacitance voltage and the magnitude of current, and when delta V is calculated dc <0，i a <i c Or DeltaV dc >0、i a <i c When sectors n=1 and n=3 are used, and other sectors n=2 and n=5 are used. The balance of the midpoint potential is adjusted using a fuzzy optimized time division factor when located in other areas. When the capacitance voltage DeltaV dc >And when 0, selecting the action time of the positive small vector as (1- < DEG > C/2, and selecting the action time of the negative small vector as (1 + < DEG > C/2). V (V) dc <And when 0, selecting the action time of the positive small vector as (1 + [ delta ] C/2, and the action time of the negative small vector as (1- [ delta ] C/2). The balancing factor ≡is obtained by the fuzzy controller.
In an optional embodiment, the controlling the on-off inversion of the bridge arm switching tube of the three-phase three-level inverter circuit includes:
bridge arm switch tube S with j phase j1 、S j2 、S j3 、S j4 The switch states of (a) are q respectively j1 、q j2 、q j3 、q j4 Then switch state q j1 、q j3 Complementary, switch state q j2 、q j4 Complementation;
capacitance C 2 The voltage V generated C2 And (V) C1 +V C2 ) The result of the comparison of/2 is input into a first PI controller which outputs a distribution ratio mu, the distribution ratio mu is input into the voltage V h The controller is used for controlling the operation of the controller,voltage V h The controller obtains the input voltage V according to a preset equation h And will input voltage V h Respectively adding to three reference voltages;
by applying a DC port voltage V dc With a target voltage V dc * The result of comparison output is input into a second PI controller, and the second PI controller outputs an inverter modulation index m;
generating three reference voltages V based on an inverter modulation index m ao *、V bo *、V co *，V ao *、V bo * And V co * Regulating the voltage of a single capacitor, outputting a voltage V jo *；
Voltage V jo * And input voltage V h The three reference voltages added respectively are compared to generate a new voltage V jo_new * Will be voltage V jo_new * The positive electrode of the first comparator and the positive electrode of the second comparator are respectively input, the positive electrode of the first comparator is compared with the negative electrode of the single-phase circuit, and the switch state q is obtained by outputting j1 The positive pole of the second comparator is compared with the negative pole of the single-phase circuit, and the switch state q is obtained by output j2 ；
Based on the output, the obtained switch state q j1 、q j2 And controlling the on-off inversion of bridge arm switching tubes of the three-phase three-level inversion circuit.
In an alternative embodiment, the equation is:
the equation is:
Vjo-R 1 i j -L j (di j /d t )-v on ＝0
v on ＝(v ao +v bo +v co )/3
v jo ＝(q j1 +q j2 -1)(v c1 +v c2 )/2
V h ＝(1-2μ)V DC /2+μV max +(1-μ)V min
wherein V is C2 For the voltage generated by the capacitor C2, V C1 Voltage q generated for capacitor C1 j1 Is a j-phase bridge arm switch tube S j1 Switch state, q j2 Is a j-phase bridge arm switch tube S j2 Switch state of V jo Voltage representing j phase, R 1 i j Represents the product of the first resistance (O passing through C) and the current of the corresponding phase j, L j (di j /d t ) Representing the product of the inductance of the respective phase and the current through the coil per unit time, v on The neutral line voltage of three phases, v ao 、v bo 、v co Represents the voltages of the phase output lines of the phase A, the phase B and the phase C respectively, mu is the distribution ratio, V dc Representing the dc port voltage.
Specifically, for an inverter circuit of j phases (including a phase, B phase, and C phase), switch S a1 ,S a2 ,S a3 ,S a4 The switch states of (a) are qj respectively 1 ,qj 2 ,qj 3 ,qj 4 The switch state is qj 1 ,qj 3 Complementary, switch state qj 2 ,qj 4 Complementary, boost converters S1, and S 2 Is qj 1 ,qj 2 ；
Three-level neutral DC charging bus (C) 1 ,C 2 ) Lower half voltage V C2 And (V) C1 +V C2 ) Input comparison, the result is input with PI, the PI outputs distribution ratio mu, the distribution ratio mu inputs voltage V h Voltage V h Range is marked V according to the equation h Added to the three reference voltages. According to the modulation index m of the inverter and according to the DC interface voltage V dc Level control carries the maximum voltage and modulation index m, and V dc * Comparing the output PI results in m being increased by three times by a single sinusoidal symbol sin (wt), sin (wt-120 deg.), sin (wt +120 deg.) generating three reference voltages V ao *、V bo * AndV co *。V ao *、V bo * And V co * Regulating voltage output V of single capacitor jo *，V jo_new * And V is equal to h Added to the three reference voltage comparisons to generate a new V jo_new * V is set up jo_new * The single-phase circuits respectively input into the positive and negative poles of the first comparator C are compared to obtain a switch state qj 1 The method comprises the steps of carrying out a first treatment on the surface of the The positive pole of the second comparator is compared with the negative pole of the single-phase circuit to output and obtain a switch state qj 2 。
The present embodiment uses an inverter as shown in fig. 5, and the following equation can be obtained by applying the basic electrical law:
V jo -R 1 i j -L h (di h /d t )-v on ＝0；
wherein V is jo Voltages representing the respective phases (e.g. A-phase, B-phase, c-phase)
R 1 i j Representing the product of the first resistance (the resistance of the O-point flowing through A, B or C) and the current of the corresponding phase; l (L) 1 (di j /d t ) Representing the product of the first inductance (LA, LB or LC) and the derivative of the corresponding inductance (current through the coil per unit time); v on The neutral line voltage is three phases;
v on ＝(v ao +v bo +v co )/3
v ao 、v bo 、v co respectively represent the voltages of the output lines of the phase A, the phase B and the phase C from the middle position
v jo ＝(q j1 +q j2 -1)(v c1 +v c2 )/2
Where j is the corresponding phase, v jo Represents the corresponding phase neutral voltage; q j1 Representing the corresponding phase switch states 1, q j2 Representing the corresponding phase switch state 2, v c1 Representative capacitance C 1 Is a voltage of (2); v c2 Representative capacitance C 2 Is a voltage of (2);
DC interface voltage (V) dc ) The level control acquires the direction in which the sufficiency file m passes through PI in the following manner. m is the ratio between the adjustment signal and the bearerBetween sufficiency, and determines that the current motivational measure is to inject it into the dc interface to maintain its desired level. A single sine symbol with three times shift of 120 degrees in m produces three reference voltages vao, vbo, and vco. To adjust the voltage of each capacitor, a zero-arrangement flag vh is added to the three reference voltages. The zero alignment flag is characterized in that:
V h ＝(1-2μ)V DC /2+μV max +(1-μ)V min
wherein μ is the distribution ratio, V C2 Is a capacitor C 2 The voltage produced, V C1 Is a capacitor C 1 The voltage q generated j1 Q is the switching state of the corresponding phase bridge arm switching tube Sa1 j2 Is the switching state of the corresponding phase leg switching tube Sa 2.
In an optional embodiment, the reference voltage V corresponding to the maximum power point is obtained based on the current state data and a preset neural network model REF And reference current I REF And based on reference voltage V REF And reference current I REF Obtaining a PWM control signal for maximum power tracking, including:
acquiring current state data;
obtaining a reference voltage V corresponding to the maximum power point based on the state data and a preset neural network model REF And reference current I REF And will reference voltage V REF And reference current I REF Switching to a second duty cycle;
according to the photovoltaic input power in the current state data, calculating an error E or an error variable delta E, inputting the error E or the error variable delta E into a fuzzy controller, outputting a duty cycle delta D by the fuzzy controller, converting the duty cycle delta D into an analog quantity through a fuzzification stage, outputting a first duty cycle after the analog quantity is converted by A/D, comparing the first duty cycle with a second duty cycle, inputting a comparison result into a grid driver, generating a PWM signal, driving DC/DC conversion by the PWM signal, driving MPPT by the DC/DC conversion, and driving the output of the maximum power of the photovoltaic array by the MPPT.
In an optional embodiment, the obtaining the reference voltage V corresponding to the maximum power point based on the state data and a preset neural network model REF And reference current I REF Comprising:
inputting the current irradiance to a preset irradiance prediction model to obtain predicted irradiance; the irradiance prediction model is obtained by training through a multi-layer perception layer device based on historical irradiance;
inputting current air temperature data into a preset air temperature prediction model to obtain predicted air temperature; the air temperature prediction model is obtained by training through an Eerman network based on historical air temperature;
inputting current photovoltaic temperature data into a preset photovoltaic temperature prediction model to obtain predicted photovoltaic temperature; the photovoltaic temperature prediction model is obtained by training through an Elman network based on historical photovoltaic temperature;
inputting the current clouds into a preset clouds prediction model to obtain predicted clouds; the cloud prediction model is based on historical clouds and is obtained by training through a back propagation neural network;
inputting predicted irradiance, predicted air temperature, predicted photovoltaic temperature, predicted cloud, current wind speed, current open-circuit voltage, current short-circuit current, current photovoltaic input voltage, photovoltaic input current and current photovoltaic input power into the neural network model to obtain a reference voltage V corresponding to a maximum power point REF And reference current I REF 。
Exemplary, as shown in fig. 6, the three-phase three-level inverter circuit is based on an environmental network neural MPPT system (8 inverters 416 photovoltaic panel setsThe formed 100KW photovoltaic park is characterized in that an energy storage module, an azimuth angle of 6.5 degrees and an inclination angle of 30 degrees are connected behind a photovoltaic direct current boosting module, a radiation meter, a wind power anemometer, a thermocouple and sensors for calibrating loads, universal meters, voltages, currents and the like are used for acquiring a photovoltaic climate parameter data acquisition network system of the weather station (meanwhile, the actual electric quantity of the load measured by the electric meter 2 is used as a predicted load (not expressed in the figure) in the future, and the photovoltaic climate parameter data of the weather station in the future is subjected to the correction in the future, is not described herein, is acquired every 15 minutes and is stored for further calculation and comparison. The data for neural network training includes an input data vector; irradiance G (W/m) 2 ) Air temperature T air (C 0 ) Battery temperature T c (C 0 ) Wind speed w (meters per second), open circuit voltage V oc (V) short-circuit current I Sc (A) The clouds, the photovoltaic input voltage, the photovoltaic input current and the photovoltaic input power form a data packet;
the voltage power period is divided into two paths of output, the first path is controlled by the fuzzy logic to calculate an error E or an error variable delta E, the error E or the error variable delta E is calculated by using a fuzzy rule base table and is converted into a language variable, the FLC fuzzy controller outputs the conversion quantity of the delta D duty ratio of a power converter, the delta D duty ratio is converted into an analog quantity through a fuzzification stage, the analog quantity is subjected to A/D conversion to obtain a first duty ratio, the first duty ratio is compared with a second duty ratio, a relatively large value duty ratio is adopted to input a grid to drive PWM, the PWM drives DC/DC conversion, the DC/DC conversion drives MPPT, and the MPPT drives the output of the maximum power of the photovoltaic panel.
The photovoltaic panel maximum power (dP/du=0) output evaluation is determined by two parameters:
wherein the subscript ref identifies the reference condition (g=1000W/m 2 ；T＝25C 0 ) Andrespectively->Is the temperature coefficient of the short circuit current and the open circuit voltage.
There are three cases in which the photovoltaic panel produces an electrical power P-V curve, (1) the characteristic P-V is linear when the ratio of the voltage V to the maximum power voltage at a given temperature is less than 0.95. The power Vmpp is closely related to the incident solar irradiance. For constant solar irradiance, there is no temperature effect in the power output.
(2) When the ratio of the voltage V to the maximum power voltage at a given temperature is greater than 1.05, the characteristic P-V drops rapidly.
(3) The scheme determined when the ratio of voltage V to maximum power voltage at a given temperature is 0.95< V/Vmpp >1.05 characterizes the state of a photovoltaic panel connected to a maximum power point tracking system (MPPT), where the energy storage, converter or industrial load is adapted to produce the maximum power point.
Automatically identifying possible abnormal values from a weather station photovoltaic weather parameter database, deleting uncorrected values, performing statistical analysis and correlation analysis and normalization processing on the data, eliminating data differences and missing data from 0.95<V/Vmpp>1.05 raw database extraction T air 、G、w，T cell V OC (G,T cell ) And I SC (G，T cell ) The method comprises the steps of carrying out a first treatment on the surface of the As the input vector of the input layer, a weight gain is selected and matched at each node of the neural network, the weight gain converts any continuous function input into any expected function, the photovoltaic array is ensured to work at an MPPT point, neurons are regularly adopted at the input layer and the hidden layers to train and test in advance, and each hidden layer has 15 hidden layer neurons. The neurons of each node accept a set of weighted inputs and generate a reference value (Vref, iref or power P) output, check the output and whether there is minimal lead error, and verify whether the training performance criteria for minimal error can be accepted, if not, reselect the following parameters until a match with the set target is entered into the test The method comprises the steps of carrying out a first treatment on the surface of the The neural network is tested for versatility and generalization. If the generalization is good, recording the output of the neural network model; output neuron = 1RF f (counter-propagating), output neuron = 1SI f (multilayer sense layer), output neuron = 1TD f (polyelman), each epoch=2000, learning rate=0.01 (back propagation), learning rate=0.1 (multi-layer perceptron), learning rate=0.1 (elman), momentum coefficient=0.9 (back propagation), threshold=1 (multi-layer perceptron), threshold=1 (elman).
The irradiance prediction model adopts a multi-layer perception layer device to transfer from input layer input to a hidden layer, the hidden layer is multiplied by synaptic weight (SV) and has hyperbolic tangent S-shaped activation function, the output of the hidden layer is transferred to an output layer, the output layer is multiplied by synaptic weight (Sw) and purelin activation function, and the multi-layer perception layer device is trained by a Leavenberg-Marquardt training method;
the cloudy (rainfall) prediction model is passed from the input layer input to the hidden layer using back propagation, the hidden layer multiplied by the synaptic weight (SV), the output of the hidden layer is passed to the output layer, the output layer multiplied by the synaptic weight (Sw) and having a hyperbolic tangent sigmoid activation function, and the back propagation is trained using a back leatenberg-Marquardt training method.
The temperature prediction model adopts an elman network and synaptic input weights (SV), and is interconnected with a hidden layer by using a hyperbolic tangent S-shaped activation function, and the output of the hidden layer is interconnected with an output layer with synaptic weights (Sw) by using a purelin activation function. The Elman network is trained by adopting a gradient descent with momentum and a self-adaptive linear back propagation training method; judging the output according to the cost, if the output is wrong, updating the cost parameter through a training algorithm and the parameter through iterative modification weight (or parameter error), and interactively selecting the output and the neuron, and adopting parameter error supervision learning training to adjust the weight;
when a group of inputs provided by the input layer do not have required outputs, unsupervised learning training is adopted, weighting coefficients are updated to obtain information of the first several layers, a next layer is provided through an intermediate hidden layer, and data mining and clustering of a multi-layer network are obtained; the topology of the neural network multi-layer perceptual network, as shown in fig. 7, using a recursive flow of signals to save and use a time sequence of events as useful information, consists of two weight layers, one hidden weight layer, two recursive function blocks and one error criterion. U is the feedback weight of the scaled input, and each signal of each recursive function is connected to a different U value.
The time parameter is scaled by u to align any weights and back-propagation is used by a gamma memory artificial neural processing element consisting of two input sources, three gamma memory blocks, three weight layers, three functional blocks, and an error criterion block. After training, for each ANN, the post-processing stage evaluates the difference between the calculated output vector and the measured output vector. The data used at this stage is not used for the training process. The neural network prediction training adopts a MATLAB platform to run on a computer with the frequency not less than 2.3GHZ and the RAM not less than 2 GB; performance assessment was performed by 6 indicators: evaluating the error qualifier verifies the performance of various artificial neural networks, such as a multi-layer perceptron of the fuzzy neural network, an improved back propagation neural network and an elman network, a parameter error supervised learning training, an unsupervised learning training, a gamma memory artificial neural processing element prediction, and the like, namely a correlation coefficient (R), a Mean Absolute Percentage Error (MAPE), a Mean Square Error (MSE), a Mean Absolute Error (MAE), a Root Mean Square Error (RMSE), a Mean Relative Error (MRE), and a time in minutes. The following mathematical equations are used for error calculation:
Where N is the total number of data samples, K' P Is the output of the object to be processed,is the average target output, +.>Is the predicted output.
An array of photovoltaic connection topology optimization using neural networks;
training experiments prove that the embodiment of the invention has the advantages that the prediction model is effective, the error is minimum, and the convergence time is shorter for tracking the maximum power Mppt of the photovoltaic array. The light irradiation prediction model multi-layer perception layer device neural network consists of 5 inputs, a single hidden layer and 14 hidden neurons, and the minimum error is realized, wherein R=1, MAPE= 2.3034e-04, MSE= 5.9791e-07, MAE= 5.0093e-04, RMSE= 7.7324e-04, MRE= 2.3034e-06 and time=1.32 minutes.
The cloudy (rain and snow) prediction model counter-propagating neural network consists of 5 inputs, a single hidden layer and 7 hidden neurons, achieving minimal error, r=1, mape=0.0158, mse=0.0042, mae=0.0320, rmse=0.0645, mre= 1.5841e-04 and time=1.24 minutes.
The temperature prediction model back propagation neural network consists of 5 inputs, a single hidden layer and 11 hidden neurons. Minimum error is achieved, map=0.1023, mse=0.0011, mae=0.0230, rmse=0.0332, mre=0.0010 and time=22 seconds.
The output layer outputs reference power P or reference current I with the required evaluation error and absolute error through the verification of the post-procedure REF And voltage V REF Reference current I REF And voltage V REF Converting into a second duty ratio, inputting the second duty ratio to perform AD conversion, comparing the second duty ratio with the first duty ratio, inputting a relatively large duty ratio into a grid electrode to drive PWM (pulse width modulation) by driving the large duty ratio, wherein the PWM is firstly controlled by a control system to directly drive three-level inverters a, b and c three-phase IGBT bridge arm switches to perform inversion according to the three-phase three-level inverter in a space vector A area from the adjusted three-level to two-level; the other path is that when the vector is positioned in the other small triangles except the area A, the time factor ≡distribution based on fuzzy control is adopted for modulation, the time factor distribution method utilizes a pair of redundant small vectors, the influence of the intermediate potential is reversely regulated, the time of the pair of small vectors is ensured to be certain, and the respective action time of the small vectors can be distributed randomly. And taking the capacitance voltage deviation e and the change rate ec of the deviation as input of a fuzzy controller, taking a balance factor ≡as fuzzy output quantity, judging and utilizing two adjacent overlapping areas to carry out voltage balance adjustment according to the difference value of the fuzzy rule and the magnitude-level capacitance voltage of the current.
With reference to fig. 9, the MPPT controlled by the MPPT controller according to the embodiment of the invention tracks the operating point rapidly and accurately under the condition of integrating the environmental factors and not changing the irradiance level, the control system sampled by the controller controls the voltage of the inverter, then generates the space vector and PWM signal for driving the IGBT, the output of the photovoltaic follows the reference thereof, the output voltage is well decoupled, the reactive power is kept constant under the variation of high dynamic performance, and the inverter can keep constant phase and line output voltage.
The PI drive fusion control of the grid electrode of the two-level-to-three-level vector fuzzy control IGBT bridge arm switch and the PWM control, as shown in the figures 11 and 12, the output voltage and current waveforms are 50Hz sine, balance and displacement, and 120 degrees of each other reveal that the inverter circuit has quite good transient and steady-state performances, the controller can accurately track the reference voltage, quickly realize steady-state values, improve the withstand voltage direct current bus voltage of the three-level inverter device by one time, and greatly reduce the waveform distortion of the inversion output current, reduce the low-order harmonic quantity and improve the quality of the output voltage by regulating the voltage balance of the direct current link capacitor. The harmonic content of the output voltage waveform was reduced to 0.29%, as shown in fig. 10. The voltage stress imposed on the switching device and the ability to operate at lower switching frequencies, which results in less dv/dt, minimizes MEI.
According to the PWM-controlled inverter device IGBT Wen Wensheng, the conduction loss is reduced, the reactive power consumption devices are reduced, ripple current around MPPT is reduced, PWM drives DC/DC conversion, the DC/DC conversion drives MPPT, and MPPT drives the output of the maximum power of the photovoltaic panel. Compared with the existing neural network training technology, the photovoltaic array power generation is controlled by adopting the neural network algorithm, so that training samples are greatly reduced, the time of months or years required by the reduction process is shortened, mismatching of the training input and the training output of the existing neural network is avoided, and a certain number of inputs and outputs can be timely interpolated or extrapolated to achieve matching. The time parameter is arbitrarily scaled by u, so that the problem that the maximum power tracking, short-time, medium-time and long-time energy storage duration of the photovoltaic is not adjustable is effectively solved. The fuzzy neural network greatly improves the tracking prediction accurate time length and accuracy, stability and reliability of the photovoltaic MPPT and the photovoltaic tracking conversion efficiency.
While the foregoing is directed to the preferred embodiments of the present invention, it will be appreciated by those skilled in the art that changes and modifications may be made without departing from the principles of the invention, such changes and modifications are also intended to be within the scope of the invention.
Claims (6)
1. A multilevel inverter system based on PWM and fuzzy neural MPPT control, comprising: the control system, the circuit breaker, the load, the photovoltaic array, the booster circuit, the three-phase three-level inverter circuit and the filter circuit are sequentially connected; wherein,,
the photovoltaic array includes: a first photovoltaic array and a second photovoltaic array;
the booster circuit includes: a first booster circuit and a second booster circuit;
the first booster circuit includes: capacitor C PV1 Inductance L 1 Switch tube S 1 And diode D 1 ；
The second booster circuit includes: capacitor C PV2 Inductance L 2 Switch tube S 2 And diode D 2 ；
The three-phase three-level inverter circuit comprises: capacitor C in series 1 And capacitor C 2 A capacitor C connected in series with 1 And capacitor C 2 The device comprises an A-phase inverter circuit, a B-phase inverter circuit and a C-phase inverter circuit which are connected in parallel;
the filter circuit includes: the first filter circuit, the second filter circuit and the third filter circuit are connected in parallel;
first photovoltaic array and capacitor C PV1 The first photovoltaic array outputs the positive electrode and the inductance L in parallel connection 1 Connected to, inductance L 1 The other end and the switching device S 1 Collector, diode D of (c) 1 Is connected with the anode of diode D 1 Cathode and capacitor C of (2) 1 One end of (C) is connected to 1 And the other end of (C) and the capacitor C 2 One end of (C) is connected to 2 And the other end of the switch tube S 2 Emitter connection of (c)
Second photovoltaic array and capacitor C PV2 The second photovoltaic array outputs the positive electrode and the inductance L in parallel connection 2 Connected to, inductance L 2 The other end and the switching device S 2 Collector, diode D of (c) 2 Is connected with the anode of diode D 2 Cathode of (d) and switching device S 1 Emitter connection of (a);
capacitor C 1 And the other end of (C) and the capacitor C 2 Is connected to the common terminal and diode D 2 Cathode of (d) and switching device S 1 Is connected with a common terminal generated by the emitter connection;
the A-phase inverter circuit includes: four bridge arm switching tubes S connected in series in sequence a1 、S a2 、S a3 、S a4 And two clamping tubes D connected in series a1 、D a2 ；
The B-phase inverter circuit includes: four bridge arm switching tubes S connected in series in sequence b1 、S b2 、S b3 、S b4 And two clamping tubes D connected in series b1 、D b2 ；
The C-phase inverter circuit includes: four bridge arm switching tubes S connected in series in sequence c1 、S c2 、S c3 、S c4 And two clamping tubes D connected in series c1 、D c2 ；
Bridge arm switch tube S a1 And bridge arm switch tube S a2 Common terminal and bridge arm switch tube S a3 And bridge arm switch tube S a4 Two clamping tubes D connected in series are connected in parallel between the common ends of the two clamping tubes a1 、D a2 Clamping tube D a1 And clamping tube D a2 Is connected to the common terminal of the capacitor C 1 And capacitor C 2 Is connected with the public end of the box body;
bridge arm switch tube S b1 And bridge arm switch tube S b2 Common terminal and bridge arm switch tube S b3 And bridge arm switch tube S b4 Two clamping tubes D connected in series are connected in parallel between the common ends of the two clamping tubes b1 、D b2 Clamping tube D b1 And clamping tube D b2 Is connected to the common terminal of the capacitor C 1 And capacitor C 2 Is connected with the public end of the box body;
bridge arm switch tube S c1 And bridge arm switch tube S c2 Common terminal and bridge arm switch tube S c3 And bridge arm switch tube S c4 Two clamping tubes D connected in series are connected in parallel between the common ends of the two clamping tubes c1 、D c2 Clamping tube D c1 And clamping tube D c2 Is connected to the common terminal of the capacitor C 1 And capacitor C 2 Is connected with the public end of the box body;
bridge arm switch tube S a2 And bridge arm switch tube S a3 The public end generated by connection is output to a first filter circuit;
bridge arm switch tube S b2 And bridge arm switch tube S b3 The public end generated by connection is output to a second filter circuit;
bridge arm switch tube S c2 And bridge arm switch tube S c3 The public end generated by connection is output to a third filter circuit;
the first output end of the filter circuit is connected with the load, and the second output end of the filter circuit is connected with the circuit breaker;
the circuit breaker is also connected with a power grid, and a public end generated by the connection of the circuit breaker and the power grid is connected with the control system;
the output of the control system is connected with the circuit breaker in a first way, the output of the control system is connected with a first controller based on the modulated three-level to two-level space vector in a second way, and the first controller is connected with a PWM controller;
The first controller is used for adjusting midpoint potential by adopting a mixed midpoint potential balance algorithm based on fuzzy optimization, adopting a seven-segment switching sequence method for the area containing the middle vector, and adopting a fuzzy optimization time factor distribution method for the area without the middle vector;
the PWM controller is used for controlling on-off inversion of bridge arm switching tubes of the three-phase three-level inverter circuit;
the controller system is also connected with an MPPT controller;
the MPPT controller is used for obtaining a reference voltage V corresponding to the maximum power point based on the current state data and a preset neural network model REF And reference current I REF And based on reference voltage V REF And reference current I REF Obtaining a PWM control signal to track the maximum power; wherein the status data comprises: irradiance, air temperature, photovoltaic temperature, wind speed, open-circuit voltage, short-circuit current, photovoltaic input voltage, photovoltaic input current and photovoltaic input power, wherein the neural network model is obtained through training according to historical state data of a system under different working conditions and corresponding output voltage and current of the corresponding system running at a maximum power point.
2. The multi-level inverter system based on PWM and fuzzy neural MPPT control of claim 1, wherein said controlling the on-off inversion of the bridge arm switching tubes of the three-phase three-level inverter circuit comprises:
Let jBridge arm switch tube S of phase j1 、S j2 、S j3 、S j4 The switch states of (a) are q respectively j1 、q j2 、q j3 、q j4 Then switch state q j1 、q j3 Complementary, switch state q j2 、q j4 Complementation;
capacitance C 2 The voltage V generated C2 And (V) C1 +V C2 ) The result of the comparison of/2 is input into a first PI controller which outputs a distribution ratio mu, the distribution ratio mu is input into the voltage V h Controller, voltage V h The controller obtains the input voltage V according to a preset equation h And will input voltage V h Respectively adding to three reference voltages;
by applying a DC port voltage V dc With a target voltage V dc * The result of comparison output is input into a second PI controller, and the second PI controller outputs an inverter modulation index m;
generating three reference voltages V based on an inverter modulation index m ao *、V bo *、V co *，V ao *、V bo * And V co * Regulating the voltage of a single capacitor, outputting a voltage V jo *；
Voltage V jo * And input voltage V h The three reference voltages added respectively are compared to generate a new voltage V jo_ * Will be voltage V jo_ * The positive electrode of the first comparator and the positive electrode of the second comparator are respectively input, the positive electrode of the first comparator is compared with the negative electrode of the single-phase circuit, and the switch state q is obtained by outputting j1 The positive pole of the second comparator is compared with the negative pole of the single-phase circuit, and the switch state q is obtained by output j2 ；
Based on the output, the obtained switch state q j1 、q j2 And controlling the on-off inversion of bridge arm switching tubes of the three-phase three-level inversion circuit.
3. The multi-level inverter system under PWM and fuzzy neural MPPT control of claim 2, wherein said equation is:
V jo -R 1 i j -L j (di j /d t )-v on ＝0
v on ＝(v ao +v bo +v co )/3
v jo ＝(q j1 +q j2 -1)(v c1 +v c2 )/2
V h ＝(1-2μ)V DC /2+μV max +(1-μ)V min
wherein V is C2 For the voltage generated by the capacitor C2, V C1 Voltage q generated for capacitor C1 j1 Is a j-phase bridge arm switch tube S j1 Switch state, q j2 Is a j-phase bridge arm switch tube S j2 Switch state of V jo Voltage representing j phase, R 1 i j Representing the product of the first resistance and the current of the corresponding phase j, L j (di j /d t ) Representing the product of the inductance of the respective phase and the current through the coil per unit time, v on The neutral line voltage of three phases, v ao 、v bo 、v co Represents the voltages of the phase output lines of the phase A, the phase B and the phase C respectively, mu is the distribution ratio, V dc Representing the dc port voltage.
4. The multilevel inverter system based on PWM and fuzzy neural MPPT control of claim 1, wherein the reference voltage V corresponding to the maximum power point is obtained based on the current state data and a preset neural network model REF And reference current I REF And based on reference voltage V REF And reference current I REF A PWM control signal is derived to perform maximum power tracking,comprising the following steps:
acquiring current state data;
Obtaining a reference voltage V corresponding to the maximum power point based on the state data and a preset neural network model REF And reference current I REF And will reference voltage V REF And reference current I REF Switching to a second duty cycle;
according to the photovoltaic input power in the current state data, calculating an error E or an error variable delta E, inputting the error E or the error variable delta E into a fuzzy controller, outputting a duty cycle delta D by the fuzzy controller, converting the duty cycle delta D into an analog quantity through a fuzzification stage, outputting a first duty cycle after the analog quantity is converted by A/D, comparing the first duty cycle with a second duty cycle, inputting a comparison result into a grid driver, generating a PWM signal, driving DC/DC conversion by the PWM signal, driving MPPT by the DC/DC conversion, and driving the output of the maximum power of the photovoltaic array by the MPPT.
5. The multilevel inverter system based on PWM and fuzzy neural MPPT control of claim 4, wherein said obtaining a reference voltage V corresponding to a maximum power point based on said state data and a predetermined neural network model REF And reference current I REF Comprising:
inputting the current irradiance to a preset irradiance prediction model to obtain predicted irradiance; the irradiance prediction model is obtained by training through a multi-layer perception layer device based on historical irradiance;
Inputting current air temperature data into a preset air temperature prediction model to obtain predicted air temperature; the air temperature prediction model is obtained by training through an Eerman network based on historical air temperature;
inputting current photovoltaic temperature data into a preset photovoltaic temperature prediction model to obtain predicted photovoltaic temperature; the photovoltaic temperature prediction model is obtained by training through an Elman network based on historical photovoltaic temperature;
inputting the current clouds into a preset clouds prediction model to obtain predicted clouds; the cloud prediction model is based on historical clouds and is obtained by training through a back propagation neural network;
inputting predicted irradiance, predicted air temperature, predicted photovoltaic temperature, predicted cloud, current wind speed, current open-circuit voltage, current short-circuit current, current photovoltaic input voltage, photovoltaic input current and current photovoltaic input power into the neural network model to obtain a reference voltage V corresponding to a maximum power point REF And reference current I REF 。
6. The multi-level inverter system under PWM and fuzzy neural MPPT control of claim 1,
The first filter circuit includes: resistor R A Inductance L A ；
The second filter circuit includes: resistor R B Inductance L B ；
The third filter circuit includes: resistor R C Inductance L C ；
Then, bridge arm switch tube S a2 And bridge arm switch tube S a3 The common terminal generated by connection is input to a resistor R A One end of the resistor R A And the other end of (2) is connected with inductance L A Is connected with one end of the connecting rod;
bridge arm switch tube S b2 And bridge arm switch tube S b3 The common terminal generated by connection is input to a resistor R B One end of the resistor R B And the other end of (2) is connected with inductance L B Is connected with one end of the connecting rod;
bridge arm switch tube S c2 And bridge arm switch tube S c3 The common terminal generated by connection is input to a resistor R C One end of the resistor R C And the other end of (2) is connected with inductance L C Is connected to one end of the connecting rod.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310444917.XA CN116470576A (en) | 2023-04-20 | 2023-04-20 | Multilevel inverter system based on PWM and fuzzy neural MPPT control |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310444917.XA CN116470576A (en) | 2023-04-20 | 2023-04-20 | Multilevel inverter system based on PWM and fuzzy neural MPPT control |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116470576A true CN116470576A (en) | 2023-07-21 |
Family
ID=87175033
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310444917.XA Pending CN116470576A (en) | 2023-04-20 | 2023-04-20 | Multilevel inverter system based on PWM and fuzzy neural MPPT control |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116470576A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117595698A (en) * | 2024-01-19 | 2024-02-23 | 安徽大学 | Inverter operation parameter prediction method, device, equipment and storage medium |
-
2023
- 2023-04-20 CN CN202310444917.XA patent/CN116470576A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117595698A (en) * | 2024-01-19 | 2024-02-23 | 安徽大学 | Inverter operation parameter prediction method, device, equipment and storage medium |
| CN117595698B (en) * | 2024-01-19 | 2024-04-09 | 安徽大学 | Inverter operation parameter prediction method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Al-Majidi et al. | A particle swarm optimisation-trained feedforward neural network for predicting the maximum power point of a photovoltaic array | |
| Messalti | A new neural networks MPPT controller for PV systems | |
| Harrag et al. | Variable step size modified P&O MPPT algorithm using GA-based hybrid offline/online PID controller | |
| Dadfar et al. | Enhanced control strategies for a hybrid battery/photovoltaic system using FGS-PID in grid-connected mode | |
| Harrag et al. | IC-based variable step size neuro-fuzzy MPPT improving PV system performances | |
| Veerachary et al. | Feedforward maximum power point tracking of PV systems using fuzzy controller | |
| Güler et al. | MPPT based model predictive control of grid connected inverter for PV systems | |
| Xu et al. | Single-phase grid-connected PV system with golden section search-based MPPT algorithm | |
| Zhang et al. | GA-RBF neural network based maximum power point tracking for grid-connected photovoltaic systems | |
| CN104965558A (en) | Photovoltaic power generation system maximum power tracking method and apparatus considering the factor of haze | |
| KR101615747B1 (en) | A efficient improvement method for maximum power point tracking of the solar inverter in case of partial shading | |
| CN102593862A (en) | Photovoltaic grid-connected inverter and control method thereof | |
| CN116470576A (en) | Multilevel inverter system based on PWM and fuzzy neural MPPT control | |
| Hu et al. | Maximum power point tracking control of a high power dc-dc converter for PV integration in MVDC distribution grids | |
| Guerra et al. | Survey on adaptative neural fuzzy inference system (ANFIS) architecture applied to photovoltaic systems | |
| CN115632394A (en) | Photovoltaic power station transient model construction and parameter identification method based on PPO algorithm | |
| CN112083753A (en) | Maximum power point tracking control method of photovoltaic grid-connected inverter | |
| Hayder et al. | Comparison of MPPT methods FLC & PSO for PV system under variable irradiance and temperature | |
| El Mezdi et al. | Nonlinear control design of grid-connected photovoltaic system through LC filter with ANN based MPPT method | |
| Abadlia et al. | Active and reactive power neurocontroller for grid-connected photovoltaic generation system | |
| Benadli et al. | A direct backstepping super-twisting algorithm controller MPPT for a standalone photovoltaic storage system: design and real-time implementation | |
| Arora et al. | AI based MPPT methods for grid connected PV systems under non linear changing solar irradiation | |
| Anu Shalini et al. | Power generation forecasting using deep learning CNN-based BILSTM technique for renewable energy systems | |
| Afghoul et al. | Comparison Study between conventional and advanced MPPT based on fuzzy logic and ANFIS for standalone system | |
| Zakri et al. | ANFIS Design Based on Prediction Models for The Photovoltaic System |
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 |