Network construction photovoltaic power generation system cooperative control method and device based on duty cycle prediction

Abstract

The application discloses a network-constructing photovoltaic power generation system cooperative control method and device based on duty cycle prediction, the method comprises: active reserve control: calculating the power deviation between the current photovoltaic power and the corresponding target load shedding power, comparing the power deviation with the power variation, and determining the power adjustment amount of the next moment according to the comparison result; the photovoltaic slope of the current moment representing the relationship between power and voltage is calculated; the power adjustment amount of the next moment, the photovoltaic slope of the current moment and the DC bus voltage are substituted into the duty cycle prediction model, and the duty cycle of the next moment is calculated; and the DC converter is controlled through the duty cycle. By adopting the technical scheme, the physical characteristic PV curve of the current photovoltaic module working point is predicted, so that the DC converter duty cycle of the next moment is obtained, instead of only relying on historical error feedback, and the dynamic response speed is improved.

Description

Technical Field

[0001] This invention relates to the fields of grid-connected control of new energy power generation and power electronics technology, and in particular to a collaborative control method and device for grid-connected photovoltaic power generation systems based on duty cycle prediction. Background Technology

[0002] With the increasing penetration rate of new energy sources, the proportion of photovoltaic power generation systems in the power system continues to rise, placing higher demands on the stable operation of the power grid. Grid-connected inverters, by simulating the characteristics of synchronous generators, can provide inertia and frequency support for the power grid, and are gradually becoming an important development direction for grid-connected photovoltaic systems.

[0003] Current photovoltaic modules and grid coordination control mostly adopt load shedding control to provide active power reserve.

[0004] Traditional load shedding control methods typically employ fixed-step perturbation or PI closed-loop regulation, gradually adjusting the DC converter's duty cycle to bring the photovoltaic output power closer to the target load shedding value. This method is essentially a step-by-step search mechanism; the controller cannot obtain the target load shedding operating point in advance and can only continuously adjust the duty cycle based on real-time power errors. Therefore, it suffers from problems such as long target operating point search time, inaccurate response to grid frequency changes, limited dynamic response speed, and susceptibility to oscillations during the load shedding process. Summary of the Invention

[0005] Purpose of the invention: This invention provides a collaborative control method and device for grid-connected photovoltaic power generation systems based on duty cycle prediction, aiming to solve the problems existing in the prior art, such as long target operating point search time, inability to accurately respond to grid frequency changes, limited dynamic response speed, and easy oscillation during load shedding.

[0006] Technical Solution: This invention provides a collaborative control method for a grid-connected photovoltaic (PV) power generation system based on duty cycle prediction, comprising: active power reserve control: calculating the power deviation between the current PV power and the corresponding target load shedding power, comparing the power deviation with the power change, and determining the power adjustment amount for the next moment based on the comparison result; calculating the PV slope at the current moment, characterizing the relationship between power and voltage, using the PV power and PV voltage at adjacent moments; the grid-connected PV power generation system includes PV modules, DC converters, inverters, and a three-phase grid, with the PV modules and DC converters connected, the DC converters and inverters connected via a DC bus, and the inverters and the three-phase grid connected via a three-phase line; substituting the power adjustment amount for the next moment, the PV slope at the current moment, and the DC bus voltage into the duty cycle prediction model, thereby calculating the duty cycle for the next moment by superimposing the duty cycle adjustment amount for the next moment with the duty cycle at the current moment; based on the next moment... The power adjustment amount and the photovoltaic slope at the current moment are used to calculate the photovoltaic voltage displacement at the next moment, and the distance between the operating point and the target point of the photovoltaic module at the next moment is determined accordingly. The duty cycle correction coefficient at the next moment is determined based on the distance, and the duty cycle adjustment amount at the next moment is corrected. The current duty cycle is then added to obtain the optimized duty cycle. When the calculated optimized duty cycle at the next moment is less than the maximum duty cycle, it is used to control the DC converter through the optimized duty cycle. The calculation of the maximum duty cycle includes: adjusting the duty cycle as a variable in the global range, recording the corresponding photovoltaic power, and taking the duty cycle corresponding to the maximum photovoltaic power as the maximum duty cycle. Grid-type control: by comparing the reference value and the actual value of the DC bus voltage, a power correction amount reflecting the energy imbalance on the DC side of the system is generated, and the difference between the power correction amount and the photovoltaic power at the current moment is used as the active power reference value of the virtual synchronous generator for grid-type control of the inverter.

[0007] Specifically, the photovoltaic slope at the current moment is obtained by calculating the ratio of the photovoltaic power difference between the current moment and the previous moment to the photovoltaic voltage difference between the current moment and the previous moment.

[0008] Specifically, when the target load reduction power corresponding to the photovoltaic power at the current moment changes, and the change exceeds the power change threshold, the ratio of the photovoltaic power to the photovoltaic voltage at the current moment is calculated to obtain the photovoltaic slope at the current moment.

[0009] Specifically, the power demand generated based on the grid frequency deviation is obtained by multiplying the deviation between the rated frequency and the grid frequency by the droop factor at the current moment, and then the target load reduction power is obtained by superimposing the reference power. The DC bus voltage at the current moment is compared with the corresponding voltage drop threshold. Based on the comparison result, it is determined whether the DC bus voltage has dropped and the droop factor at the current moment. When the DC bus voltage drops, the larger value of the enhanced support gain droop factor is selected; when the DC bus voltage does not drop, the smaller value of the basic support gain droop factor is selected.

[0010] Specifically, based on the relationship between the input voltage and output voltage of the DC converter and the duty cycle, the basic relationship formulas between the DC bus voltage and the photovoltaic voltage and the duty cycle are obtained; the basic relationship formulas with variables substituted into the next time step and the current time step are differentially processed, and the photovoltaic slope is substituted to map the photovoltaic voltage to the photovoltaic power, thus obtaining the duty cycle prediction model.

[0011] Specifically, the power adjustment amount at the next moment is used to replace the difference in photovoltaic power between the current moment and the previous moment in the duty cycle prediction model.

[0012] Specifically, the formula for calculating the optimized duty cycle is as follows: d'(k+1)=-γ(k+1)(P use (k+1) / k pv (k)V dc )+d(k), Where d'(k+1) represents the optimized duty cycle at the next time step, d(k) represents the duty cycle at the current time step, γ(k+1) represents the duty cycle correction coefficient at the next time step, and P use (k+1) represents the power adjustment amount at the next moment, k pv (k) represents the photovoltaic slope at the current moment, V dc This indicates the DC bus voltage.

[0013] Specifically, the smaller the photovoltaic voltage displacement, the smaller the distance between the operating point and the target point of the photovoltaic module, and the closer the duty cycle correction coefficient is to 1; the larger the photovoltaic voltage displacement, the larger the distance between the operating point and the target point of the photovoltaic module, and the closer the duty cycle correction coefficient is to 0.

[0014] Specifically, it also includes: when the optimized duty cycle for the next calculated time step is greater than or equal to the maximum duty cycle, it is used to control the photovoltaic module through the maximum duty cycle.

[0015] This invention also provides a collaborative control device for a grid-connected photovoltaic power generation system based on duty cycle prediction, comprising: an active power reserve unit and a grid-connected control unit, wherein: the active power reserve unit is used to calculate the power deviation between the photovoltaic power at the current moment and the corresponding target load reduction power, compare the power deviation with the power change, and determine the power adjustment amount at the next moment based on the comparison result; and calculate the photovoltaic slope at the current moment, which characterizes the relationship between power and voltage, using the photovoltaic power and photovoltaic voltage at adjacent moments; the grid-connected photovoltaic power generation system includes photovoltaic modules, DC converters, inverters, and a three-phase power grid, wherein the photovoltaic modules and DC converters are connected, the DC converters and inverters are connected through a DC bus, and the inverters and the three-phase power grid are connected through a three-phase line; the power adjustment amount at the next moment, as well as the photovoltaic slope and DC bus voltage at the current moment, are substituted into the duty cycle prediction model, thereby calculating the duty cycle at the next moment by superimposing the duty cycle adjustment amount at the next moment with the duty cycle at the current moment; Based on the power adjustment amount at the next moment and the photovoltaic slope at the current moment, the photovoltaic voltage displacement at the next moment is calculated, and the distance between the operating point and the target point of the photovoltaic module at the next moment is determined accordingly. The duty cycle correction coefficient at the next moment is determined based on the distance, and the duty cycle adjustment amount at the next moment is corrected. The current duty cycle is then added to obtain the optimized duty cycle. When the calculated optimized duty cycle at the next moment is less than the maximum duty cycle, it is used to control the DC converter through the optimized duty cycle. The calculation of the maximum duty cycle includes: adjusting the duty cycle as a variable in the global range, recording the corresponding photovoltaic power, and taking the duty cycle corresponding to the maximum photovoltaic power as the maximum duty cycle. The grid-type control unit is used to generate a power correction amount reflecting the energy imbalance on the DC side of the system by comparing the reference value and the actual value of the DC bus voltage, and the difference between the power correction amount and the photovoltaic power at the current moment is used as the active power reference value of the virtual synchronous generator for grid-type control of the inverter.

[0016] Beneficial Effects: Compared with existing technologies, this invention has the following significant advantages: It predicts the DC converter duty cycle for the next moment based on the physical characteristics of the PV curve at the current operating point of the photovoltaic module, rather than relying solely on historical error feedback, thus improving dynamic response speed. Furthermore, a correction scheme is introduced in the calculation of the duty cycle for future moments to avoid system oscillations during load shedding. Furthermore, in the calculation of the duty cycle (related to the target load shedding power), grid frequency changes and DC bus voltage drops (related to the droop coefficient) are taken into consideration, enabling the calculated duty cycle to accurately and effectively respond to grid frequency changes and DC bus voltage drops, achieving coordinated support for both grid frequency and DC bus voltage. Furthermore, based on coordinated grid-type control of the inverter on the DC bus side and the photovoltaic side, the AC side of the system can respond to changes on the DC side and the photovoltaic side, improving the overall stability of the system operation. Attached Figure Description

[0017] Figure 1 A topology diagram of the grid-type photovoltaic power generation system and control method provided by the present invention; Figure 2 A waveform comparison diagram of inverter output active power and DC bus voltage provided by the present invention; Figure 3 A waveform comparison diagram of photovoltaic module voltage and grid-side frequency provided for this invention. Detailed Implementation

[0018] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0019] See Figure 1 This is a topology diagram of the grid-type photovoltaic power generation system and control method provided by the present invention.

[0020] In practical implementation, the grid-type photovoltaic power generation system includes photovoltaic modules (photovoltaic units S1 to S3), DC (DC-DC) converters, inverters (DC-AC), and a three-phase grid. The photovoltaic modules and DC converters are connected, the DC converters and inverters are connected through a DC bus, and the inverters and the three-phase grid are connected through three-phase lines. An LCL-type filter circuit structure is also installed on the three-phase lines.

[0021] In practice, active power reserve is usually used to provide active power compensation when the grid frequency drops or the DC bus voltage drops.

[0022] In practical implementation, a DC-DC converter controls the conversion relationship between the input voltage and the output voltage by adjusting the ratio of the on and off times (i.e., the duty cycle) of its switching transistors, thereby achieving regulation of the operating point of the photovoltaic array.

[0023] In practical implementation, for this grid-type photovoltaic power generation system, active power reserve is mainly based on active power compensation from the photovoltaic power output of the photovoltaic modules. The output of the photovoltaic modules is related to the duty cycle of the DC converter. Therefore, the active power reserve control provided by this invention is mainly based on the current operating point of the photovoltaic modules, responding to changes on the grid side and DC side, predicting the duty cycle of the DC converter at the next moment, and simultaneously increasing or maintaining the output power of the photovoltaic modules to provide active power support for the grid side and DC side, ensuring the stable operation of the system.

[0024] In this embodiment of the invention, the photovoltaic power at the current moment is calculated (this can be done by collecting the voltage V of the photovoltaic module). pv and current I pv The power deviation between the calculated target load reduction power and the corresponding target load reduction power is compared with the power change, and the power adjustment amount for the next moment is determined based on the comparison result.

[0025] In this embodiment of the invention, the calculation of the target load reduction power includes: multiplying the deviation between the rated frequency and the grid frequency by the droop coefficient at the current moment to obtain the power demand generated based on the grid frequency deviation, and then superimposing the reference power to obtain the target load reduction power; comparing the DC bus voltage at the current moment with the corresponding voltage drop threshold, and determining whether the DC bus voltage has dropped and the droop coefficient at the current moment based on the comparison result; when the DC bus voltage drops, selecting the larger value of the enhanced support gain droop coefficient; when the DC bus voltage does not drop, selecting the smaller value of the basic support gain droop coefficient.

[0026] In practical implementation, the formula for calculating the target load reduction power is as follows: P ref =P0-K(f n -f), Among them, P ref P0 represents the target load reduction power, K represents the baseline power, and f represents the droop coefficient. n K(f) and f represent the rated frequency and the mains frequency, respectively. n The -f) term represents the power demand generated based on the grid frequency deviation.

[0027] In specific implementation, the droop coefficient K is determined as follows: when the DC bus voltage is greater than or equal to the voltage drop threshold (when the DC bus voltage has not dropped), K is the basic support gain droop coefficient; when the DC bus voltage is less than the voltage drop threshold (when the DC bus voltage has dropped), K is the enhanced support gain droop coefficient. The basic support gain droop coefficient is less than the enhanced support gain droop coefficient. The specific value can be set according to the actual application scenario.

[0028] In practical implementation, the target load reduction power is one of the important factors in the calculation of the DC converter duty cycle. The target load reduction power is related to the grid frequency and DC bus voltage, so that the calculated duty cycle can accurately and effectively respond to changes in grid frequency and DC bus voltage drops, and achieve coordinated support for grid frequency and DC bus voltage.

[0029] In practical implementation, the power deviation e p The calculation method is e p =P ref -P pv (k), where P pv (k) represents the photovoltaic power at time k.

[0030] In practical implementation, a proportional coefficient Ks is introduced for calculating the power adjustment at the next moment, specifically as follows: Kse p >P max At time k+1, the power adjustment P use (k+1)=P max ,|Kse p |≤P max At time k+1, the power adjustment P use (k+1)=Kse p Kse p <-P max At time k+1, the power adjustment P use (k+1)=-P max .

[0031] In practical implementation, the calculation logic for the power regulation at the next moment is to select Kse. p (The value of Ks can be determined based on the actual application scenario) serves as the basic power regulation amount. When the basic power regulation amount is less than the maximum power change in a single operation (|Kse) p |≤P max ), output Kse p As the power adjustment amount at the next moment, with e p continuously decreasing, Kse p It also decreases synchronously, and the adjustment amount smoothly follows the error response. When the base power adjustment amount is greater than the single maximum power change amount (|Kse) p |>P max The output is limited to ±P. max .

[0032] In this embodiment of the invention, the photovoltaic slope at the current moment, which characterizes the relationship between power and voltage, is calculated by using photovoltaic power and photovoltaic voltage at adjacent moments.

[0033] In this embodiment of the invention, the photovoltaic slope at the current moment is obtained by calculating the ratio of the photovoltaic power difference between the current moment and the previous moment to the photovoltaic voltage difference between the current moment and the previous moment.

[0034] In practical implementation, the formula for calculating the photovoltaic slope is: k pv (k)=(P pv (k)-P pv (k-1)) / (V pv (k)-V pv (k-1)), Where, k pv (k) represents the photovoltaic slope at time k, P pv (k) and P pv (k-1) represent the photovoltaic power at time k and time k-1, respectively, V pv (k) and V pv (k-1) represents the photovoltaic voltage at time k and time k-1, respectively.

[0035] In practice, the above calculation method is based on the fact that adjacent moments are relatively close in time and the slope changes gently, which can approximately satisfy local linearity.

[0036] In this embodiment of the invention, when the target load reduction power corresponding to the photovoltaic power at the current moment changes and the change exceeds the power change threshold, the ratio of the photovoltaic power to the photovoltaic voltage at the current moment is calculated to obtain the photovoltaic slope at the current moment.

[0037] In practical implementation, photovoltaic (PV) characteristics exhibit strong local nonlinearity, and the dynamic characteristics differ significantly across different operating regions. When the target load shedding power changes, and the change exceeds the power change threshold, it indicates that historical slope information is no longer applicable for prediction in the new operating region. Therefore, the current P and V values ​​are used to reset the slope and determine the correct direction. The specific calculation method is as follows: When the target load reduction power changes at time k, and the change exceeds the power change threshold (relative to the target load reduction power at time k-1), k pv (k)=-|P pv (k) / V pv (k)| means replacing the original photovoltaic slope with the newly calculated one, and doing so in subsequent calculations. This is a correction scheme introduced in the calculation of the duty cycle at future times to avoid the problem of system oscillations during load shedding.

[0038] In this embodiment of the invention, the power adjustment amount at the next moment, as well as the photovoltaic slope and DC bus voltage at the current moment, are substituted into the duty cycle prediction model. Thus, the duty cycle at the next moment is calculated by superimposing the duty cycle adjustment amount at the next moment with the duty cycle at the current moment.

[0039] In this embodiment of the invention, based on the relationship between the input voltage and output voltage of the DC converter and the duty cycle, the basic relationship formulas between the DC bus voltage and the photovoltaic voltage and the duty cycle are obtained; the basic relationship formulas with variables substituted into the next time moment and the current time moment are differentially processed, and the photovoltaic slope is substituted to map the photovoltaic voltage to the photovoltaic power, thus obtaining the duty cycle prediction model.

[0040] In practical implementation, the relationship between the input voltage and output voltage of the DC converter and the duty cycle is as follows: V o =V in / (1-d), Among them, V o and V in These represent the output voltage and input voltage of the DC converter, respectively. The input voltage V in Similarly, the photovoltaic voltage V pv Output voltage V o Similarly, the DC bus voltage V dc d represents the duty cycle of the DC converter.

[0041] In practical implementation, based on the above, the fundamental relationship formulas between DC bus voltage, photovoltaic voltage, and duty cycle can be obtained: V pv =(1-d)V dc .

[0042] In practical implementation, let the DC bus voltage V be... dc If the voltage remains constant over two adjacent control cycles, then by differencing the equations at times k and k+1, we can obtain the relationship between the change in photovoltaic voltage and the change in duty cycle: V pv (k+1)-V pv (k)=-(d(k+1)-d(k))V dc , Among them, V pv (k+1) and V pv (k) represents the photovoltaic voltage at time k+1 and k, respectively, and d(k+1) and d(k) represent the duty cycle at time k+1 and k, respectively.

[0043] Substituting the photovoltaic slope into the mapping from photovoltaic voltage to photovoltaic power, we get: P pv (k+1)-P pv(k)=-k pv (k)(d(k+1)-d(k))V dc , Furthermore, the duty cycle prediction model is obtained: d(k+1)=-(P pv (k+1)-P pv (k)) / (k pv (k)V dc )+d(k), Among them, P pv (k+1) and P pv (k) represents the photovoltaic power at times k+1 and k, respectively. pv (k) represents the photovoltaic slope at time k.

[0044] In this embodiment of the invention, the power adjustment amount at the next moment is used to replace the photovoltaic power difference between the current moment and the previous moment in the duty cycle prediction model.

[0045] In practical implementation, in the duty cycle prediction model -(P pv (k+1)-P pv (k)) / (k pv (k)V dc ) is the duty cycle adjustment amount for the next time step (at time k). The duty cycle d(k) of the current time step (at time k) is added to obtain the duty cycle for the next time step.

[0046] In practical implementation, the corresponding term in the duty cycle prediction model is P. pv (k+1)-P pv (k), which is the difference in photovoltaic power between the next time step and the current time step, P pv (k+1) also represents the target power, the adjustment target of photovoltaic power, that is, the target load reduction power P. ref Therefore, theoretically, the power deviation e should be substituted. p That is, e p =P ref -P pv (k). However, if we use e directly... p Substituting local slopes into the duty cycle prediction model will lead to inaccurate predictions and excessive errors. Therefore, e p The power increment is decomposed into multiple segments, namely the power adjustment P at the next moment calculated above. use Substituting (k+1) into the duty cycle prediction model can make the operating point of the photovoltaic module gradually approach the target load reduction point.

[0047] In this embodiment of the invention, the photovoltaic voltage displacement at the next moment is calculated based on the power adjustment amount at the next moment and the photovoltaic slope at the current moment, and the distance between the operating point and the target point of the photovoltaic module at the next moment is determined accordingly. The duty cycle correction coefficient at the next moment is determined based on the magnitude of the distance.

[0048] In practical implementation, the photovoltaic voltage displacement V at the next moment p The calculation is as follows: V p =P use (k+1) / k pv (k).

[0049] In this embodiment of the invention, the smaller the photovoltaic voltage displacement, the smaller the distance between the operating point and the target point of the photovoltaic module, and the closer the duty cycle correction coefficient is to 1; the larger the photovoltaic voltage displacement, the larger the distance between the operating point and the target point of the photovoltaic module, and the closer the duty cycle correction coefficient is to 0.

[0050] In practical implementation, the duty cycle correction factor γ is calculated as follows: γ(k+1)=1 / (1+αV p ), Where α represents the change sensitivity coefficient, and γ(k+1) represents the duty cycle correction coefficient at time k+1.

[0051] In practical implementation, when the photovoltaic voltage displacement V p A smaller value indicates that the operating point (the point on the PV / IV curve) is closer to the target point (the operating point corresponding to the target load reduction power), and the corresponding photovoltaic slope k is... pv The local area covered by the curve is relatively small, so the photovoltaic slope calculated based on historical photovoltaic information, and the corresponding photovoltaic power prediction results, can be largely adopted, at which point γ is close to 1; when the photovoltaic voltage displacement V p When the value increases, it indicates that the distance between the working point and the target point increases, and the photovoltaic slope k... pv The covered curve has a relatively large local range. Due to the nonlinear characteristics of photovoltaic (PV) properties, the greater the voltage change, the higher the risk that the PV slope calculated based on historical PV information will accurately describe the characteristics of the future operating point. Therefore, by reducing γ to dampen the duty cycle adjustment, the over-adjustment phenomenon that may be caused by model errors is reduced, thereby improving the stability and robustness of the load shedding process. This is a correction scheme introduced in the calculation of the duty cycle at future times to avoid the problem of system oscillations that are prone to occur during the load shedding process.

[0052] In this embodiment of the invention, the duty cycle adjustment amount at the next moment is corrected, and the duty cycle at the current moment is superimposed to obtain the optimized duty cycle.

[0053] In this embodiment of the invention, the formula for calculating the optimized duty cycle is as follows: d'(k+1)=-γ(k+1)(P use (k+1) / k pv (k)V dc )+d(k), Where d'(k+1) represents the optimized duty cycle at the next time step, d(k) represents the duty cycle at the current time step, γ(k+1) represents the duty cycle correction coefficient at the next time step, and P use (k+1) represents the power adjustment amount at the next moment, k pv (k) represents the photovoltaic slope at the current moment, V dc This indicates the DC bus voltage.

[0054] In practical implementation, γ(k+1) is relative to -(P use (k+1) / k pv (k)V dc The item ) needs to be corrected.

[0055] In this embodiment of the invention, when the calculated optimized duty cycle for the next moment is less than the maximum duty cycle, it is used to control the DC converter by optimizing the duty cycle (generated by a PWM signal).

[0056] In this embodiment of the invention, when the optimized duty cycle at the next calculated moment is greater than or equal to the maximum duty cycle, it is used to control the photovoltaic module through the maximum duty cycle.

[0057] In practical implementation, by setting duty cycle constraints, i.e., the duty cycle of the applied DC converter is less than or equal to the maximum duty cycle, the photovoltaic array is always operated in the load reduction region to the right of the maximum power point. This effectively avoids the problem of photovoltaics accidentally entering local extreme points under local shading conditions, weakens power oscillations during load reduction switching, and ensures stable active power reserve capacity and high steady-state control accuracy.

[0058] In this embodiment of the invention, the calculation of the maximum duty cycle includes: adjusting the duty cycle as a variable in the global range, recording the corresponding photovoltaic power, taking the duty cycle corresponding to the maximum photovoltaic power as the maximum duty cycle, and tracking the maximum power point Mppt.

[0059] In practical implementation, "global range" refers to a relatively broad range of duty cycle values ​​to cover various operating conditions as much as possible and avoid getting trapped in local extrema. For example, within the duty cycle range of [0.1, 0.95], the algorithm iterates through N=25 equally spaced sampling points. During each sampling period, photovoltaic voltage and current are collected in real time, and the power value at each point is calculated and recorded. After the scan is completed, the full-range power data is compared to identify the global maximum power and the corresponding duty cycle command. Once the global maximum power is locked, the algorithm automatically switches from full-range scanning to Incremental Conductivity (INC) fine-tuning mode, using the global maximum power as the starting search point for the INC algorithm to perform a fine search.

[0060] In this embodiment of the invention, the grid-type control of the grid-connected photovoltaic power generation system includes: generating a power correction amount that reflects the energy imbalance on the DC side of the system by comparing the reference value and the actual value of the DC bus voltage, and using the difference between the power correction amount and the photovoltaic power at the current moment as the active power reference value of the virtual synchronous generator (VSG) for grid-type control of the inverter.

[0061] In practical implementation, the active power reference value P of the virtual synchronous generator mref The calculation method is as follows: P mref =P pv -△P, Among them, P pv ΔP represents the photovoltaic power, and ΔP represents the power correction amount obtained by PI control of the DC bus voltage, which reflects the energy imbalance on the DC side of the system.

[0062] In practical implementation, in the voltage and current dual closed-loop control, the power at the grid connection point is collected (which can be calculated by collecting the voltage and current at the grid connection point), and the corresponding reactive power Q and active power P are calculated. e Feedback control is based on active power and the corresponding active power reference value.

[0063] In practical implementation, the inverter is controlled in a coordinated grid configuration based on the DC bus side and the photovoltaic side, so that the AC side of the system can respond to changes in the DC side and the photovoltaic side, thereby improving the overall stability of the system operation.

[0064] See Figure 2 The diagram shows a waveform comparison of the inverter output active power and DC bus voltage provided by this invention.

[0065] exist Figure 2 In the middle (including two sub-graphs), the network-side frequency decreases when the setting is 1.5 to 2.5 seconds.

[0066] like Figure 2The inverter output active power shown in the upper sub-diagram, under the complex power-voltage curve characteristics caused by external factors such as local shading, employs the INC (Incremental Conductance Method) scheme, as shown by the blue waveform in the figure. The inverter remains trapped at a local power peak point (approximately 49.9kW) and cannot identify the global maximum power location. Furthermore, due to the lack of active power reserve, the output power remains unchanged when the frequency begins to drop after 1.5 seconds, failing to provide any active power compensation to the grid. The red waveform represents the active power waveform of the active power reserve method with a fixed droop coefficient K. It can be seen that, compared to the blue waveform, it can generate more active power for frequency support during frequency drops, but with a fixed droop coefficient, the power increase is constrained by constant parameters, resulting in limited adjustment margin. The black waveform represents the standby power support control scheme proposed in this invention, based on adaptive setting of the droop coefficient using DC bus voltage. This scheme can moderately increase the active power output within its own standby capacity range, with a higher power increase amplitude than the fixed droop coefficient scheme, providing stronger power support.

[0067] like Figure 2 The lower sub-graph shows the DC bus voltage, with a reference value of 800V. The blue waveform represents the INC (Incremental Conductivity Method) scheme without backup power support. Under grid frequency sag disturbances, the DC bus voltage experienced a severe transient sag, with the lowest voltage dropping to approximately 778V, representing the largest sag depth. Significant overshoot accompanied the voltage recovery process, resulting in the worst DC-side stability. The red waveform represents the control scheme with backup power support but still using a fixed droop coefficient. Compared to the unsupported scheme, the increased backup power effectively mitigated the voltage sag, raising the lowest voltage point to approximately 786V and reducing the sag depth. However, due to the limitations of the fixed coefficient, the transient support effect was still insufficient, and significant fluctuations occurred during the voltage recovery process. The black waveform represents the standby power support control scheme proposed in this invention, which is based on adaptive adjustment of the droop coefficient using DC bus voltage. With the same standby power configuration as the red curve, this scheme dynamically adjusts the droop coefficient K by sensing the DC bus voltage status in real time. When the DC bus voltage is greater than 795V, a smaller coefficient is used to ensure steady-state smoothness. When the DC bus voltage is less than 795V, the coefficient is automatically increased to strengthen active power generation support. The lowest point of the voltage transient drop on the black curve is raised to about 790V, and the drop depth is further reduced by about 4V compared with the traditional scheme with a fixed droop coefficient. The transient support effect is significantly enhanced, and the voltage recovery process is smooth without obvious overshoot.

[0068] See Figure 3 This is a waveform comparison diagram of the photovoltaic module voltage and grid-side frequency provided by the present invention.

[0069] exist Figure 3 In the middle (including two sub-graphs), the network-side frequency decreases when the setting is 1.5 to 2.5 seconds.

[0070] like Figure 3 The upper sub-figure shows the photovoltaic module voltage. The decrease in photovoltaic voltage provided by the method of this invention when the frequency changes proves that when the grid-side frequency drops, the photovoltaic operating point shifts to the left along the PV curve, thus increasing the output power. In contrast, the photovoltaic bus voltage obtained by the incremental conductance method remains at the voltage corresponding to the local peak.

[0071] like Figure 3 The grid-side frequency shown in the lower sub-figure exhibits significant frequency oscillations in the incremental conductance method, while the frequency change of the method provided by this invention is smoother and the rate of frequency change is smaller, thus achieving effective frequency support.

[0072] This invention also provides a collaborative control device for a grid-connected photovoltaic power generation system based on duty cycle prediction, comprising: an active power reserve unit and a grid-connected control unit, wherein: the active power reserve unit is used to calculate the power deviation between the photovoltaic power at the current moment and the corresponding target load reduction power, compare the power deviation with the power change, and determine the power adjustment amount at the next moment based on the comparison result; and calculate the photovoltaic slope at the current moment, which characterizes the relationship between power and voltage, using the photovoltaic power and photovoltaic voltage at adjacent moments; the grid-connected photovoltaic power generation system includes photovoltaic modules, DC converters, inverters, and a three-phase power grid, wherein the photovoltaic modules and DC converters are connected, the DC converters and inverters are connected through a DC bus, and the inverters and the three-phase power grid are connected through a three-phase line; the power adjustment amount at the next moment, as well as the photovoltaic slope and DC bus voltage at the current moment, are substituted into the duty cycle prediction model, thereby calculating the duty cycle at the next moment by superimposing the duty cycle adjustment amount at the next moment with the duty cycle at the current moment; Based on the power adjustment amount at the next moment and the photovoltaic slope at the current moment, the photovoltaic voltage displacement at the next moment is calculated, and the distance between the operating point and the target point of the photovoltaic module at the next moment is determined accordingly. The duty cycle correction coefficient at the next moment is determined based on the distance, and the duty cycle adjustment amount at the next moment is corrected. The current duty cycle is then added to obtain the optimized duty cycle. When the calculated optimized duty cycle at the next moment is less than the maximum duty cycle, it is used to control the DC converter through the optimized duty cycle. The calculation of the maximum duty cycle includes: adjusting the duty cycle as a variable in the global range, recording the corresponding photovoltaic power, and taking the duty cycle corresponding to the maximum photovoltaic power as the maximum duty cycle. The grid-type control unit is used to generate a power correction amount reflecting the energy imbalance on the DC side of the system by comparing the reference value and the actual value of the DC bus voltage, and the difference between the power correction amount and the photovoltaic power at the current moment is used as the active power reference value of the virtual synchronous generator for grid-type control of the inverter.

[0073] In specific implementation, the methods, steps or functions executed by the execution unit of the grid-connected photovoltaic power generation system collaborative control device based on duty cycle prediction provided by the present invention can refer to the grid-connected photovoltaic power generation system collaborative control method based on duty cycle prediction provided by the present invention.

Claims

1. A collaborative control method for a grid-connected photovoltaic power generation system based on duty cycle prediction, characterized in that, include: Active power reserve control: The power deviation between the current photovoltaic power and the corresponding target load reduction power is calculated, and the power deviation is compared with the power change. Based on the comparison result, the power adjustment amount for the next moment is determined. The photovoltaic slope at the current moment, which characterizes the relationship between power and voltage, is calculated using the photovoltaic power and photovoltaic voltage at adjacent moments. The grid-type photovoltaic power generation system includes photovoltaic modules, DC converters, inverters, and a three-phase power grid. The photovoltaic modules and DC converters are connected, the DC converters and inverters are connected through a DC bus, and the inverters and the three-phase power grid are connected through a three-phase line. Substitute the power adjustment amount at the next moment, along with the photovoltaic slope and DC bus voltage at the current moment, into the duty cycle prediction model. Then, by superimposing the duty cycle adjustment amount at the next moment with the duty cycle at the current moment, the duty cycle at the next moment can be calculated. Based on the power adjustment amount at the next moment and the photovoltaic slope at the current moment, the photovoltaic voltage displacement at the next moment is calculated, and the distance between the operating point and the target point of the photovoltaic module at the next moment is determined accordingly. The duty cycle correction coefficient at the next moment is determined based on the magnitude of the distance, and the duty cycle adjustment amount at the next moment is corrected. The current duty cycle is then added to obtain the optimized duty cycle. When the calculated optimized duty cycle for the next time step is less than the maximum duty cycle, it is used to control the DC converter by optimizing the duty cycle. The calculation of the maximum duty cycle includes: adjusting the duty cycle as a variable in the global scope, recording the corresponding photovoltaic power, and taking the duty cycle corresponding to the maximum photovoltaic power as the maximum duty cycle; Network-based control: By comparing the reference value and the actual value of the DC bus voltage, a power correction amount reflecting the energy imbalance on the DC side of the system is generated. The difference between the power correction amount and the photovoltaic power at the current moment is used as the active power reference value of the virtual synchronous generator for grid-type control of the inverter.

2. The collaborative control method for grid-connected photovoltaic power generation systems based on duty cycle prediction according to claim 1, characterized in that, The calculation of the photovoltaic slope at the current moment, which characterizes the relationship between power and voltage, using photovoltaic power and photovoltaic voltage at adjacent moments, includes: The photovoltaic slope at the current moment is obtained by calculating the ratio of the difference in photovoltaic power between the current moment and the previous moment to the difference in photovoltaic voltage between the current moment and the previous moment.

3. The collaborative control method for grid-connected photovoltaic power generation systems based on duty cycle prediction according to claim 2, characterized in that, The process of obtaining the photovoltaic slope at the current moment includes: When the target load reduction power corresponding to the photovoltaic power at the current moment changes, and the change exceeds the power change threshold, the ratio of the photovoltaic power to the photovoltaic voltage at the current moment is calculated to obtain the photovoltaic slope at the current moment.

4. The collaborative control method for grid-connected photovoltaic power generation systems based on duty cycle prediction according to claim 3, characterized in that, The target load reduction power includes: The power demand generated based on the grid frequency deviation is obtained by multiplying the deviation between the rated frequency and the grid frequency by the droop factor at the current moment, and then superimposing the reference power to obtain the target load reduction power. The DC bus voltage at the current moment is compared with the corresponding voltage drop threshold. Based on the comparison result, it is determined whether the DC bus voltage has dropped and the droop factor at the current moment. When the DC bus voltage drops, the larger value of the enhanced support gain droop factor is selected; when the DC bus voltage does not drop, the smaller value of the basic support gain droop factor is selected.

5. The collaborative control method for grid-connected photovoltaic power generation systems based on duty cycle prediction according to claim 1, characterized in that, The calculation process of the duty cycle prediction model includes: Based on the relationship between the input voltage and output voltage of a DC converter and its duty cycle, the basic relationship formulas between the DC bus voltage and the photovoltaic voltage and their duty cycle are obtained. The basic relationship formulas of the variables substituted into the next time step and the current time step are differentially expressed, and the photovoltaic slope is substituted to map the photovoltaic voltage to the photovoltaic power, thus obtaining the duty cycle prediction model.

6. The collaborative control method for grid-connected photovoltaic power generation systems based on duty cycle prediction according to claim 5, characterized in that, The duty cycle prediction model also includes: Replace the photovoltaic power difference between the current time and the previous time in the duty cycle prediction model with the power adjustment amount at the next time step.

7. The collaborative control method for grid-connected photovoltaic power generation systems based on duty cycle prediction according to claim 6, characterized in that, The formula for calculating the optimized duty cycle is as follows: d'(k+1)=-γ(k+1)(P use (k+1) / k pv (k)V dc )+d(k), Where d'(k+1) represents the optimized duty cycle at the next time step, d(k) represents the duty cycle at the current time step, γ(k+1) represents the duty cycle correction coefficient at the next time step, and P use (k+1) represents the power adjustment amount at the next moment, k pv (k) represents the photovoltaic slope at the current moment, V dc This indicates the DC bus voltage.

8. The collaborative control method for grid-connected photovoltaic power generation systems based on duty cycle prediction according to claim 6, characterized in that, The calculation yields the photovoltaic voltage displacement at the next moment, and based on this, the distance between the operating point and the target point of the photovoltaic module at the next moment is determined. The duty cycle correction coefficient for the next moment is then determined based on the magnitude of this distance, including: The smaller the photovoltaic voltage displacement, the smaller the distance between the operating point and the target point of the photovoltaic module, and the closer the duty cycle correction coefficient is to 1; the larger the photovoltaic voltage displacement, the larger the distance between the operating point and the target point of the photovoltaic module, and the closer the duty cycle correction coefficient is to 0.

9. The collaborative control method for grid-connected photovoltaic power generation systems based on duty cycle prediction according to claim 1, characterized in that, Also includes: When the calculated optimized duty cycle for the next time step is greater than or equal to the maximum duty cycle, it is used to control the photovoltaic module using the maximum duty cycle.

10. A collaborative control device for a grid-connected photovoltaic power generation system based on duty cycle prediction, characterized in that, Includes: active power reserve units and network-type control units, wherein: The active power reserve unit is used to calculate the power deviation between the current photovoltaic power and the corresponding target load reduction power, compare the power deviation with the power change, and determine the power adjustment amount for the next moment based on the comparison result; it also calculates the photovoltaic slope at the current moment, which characterizes the relationship between power and voltage, using the photovoltaic power and photovoltaic voltage at adjacent moments; the grid-type photovoltaic power generation system includes photovoltaic modules, DC converters, inverters, and a three-phase power grid. The photovoltaic modules and DC converters are connected, the DC converters and inverters are connected through a DC bus, and the inverters and the three-phase power grid are connected through a three-phase line; the power adjustment amount for the next moment, as well as the photovoltaic slope and DC bus voltage at the current moment, are substituted into the duty cycle prediction model, thereby adjusting the power based on the duty cycle at the next moment. The duty cycle at the current moment is superimposed to calculate the duty cycle at the next moment. Based on the power adjustment at the next moment and the photovoltaic slope at the current moment, the photovoltaic voltage displacement at the next moment is calculated, and the distance between the operating point and the target point of the photovoltaic module at the next moment is determined accordingly. The duty cycle correction coefficient at the next moment is determined based on the distance, and the duty cycle adjustment at the next moment is corrected. The current duty cycle is then superimposed to obtain the optimized duty cycle. When the calculated optimized duty cycle at the next moment is less than the maximum duty cycle, it is used to control the DC converter through the optimized duty cycle. The calculation of the maximum duty cycle includes: adjusting the duty cycle as a variable in the global range, recording the corresponding photovoltaic power, and taking the duty cycle corresponding to the maximum photovoltaic power as the maximum duty cycle. The grid-type control unit is used to generate a power correction amount that reflects the energy imbalance on the DC side of the system by comparing the reference value and the actual value of the DC bus voltage, and uses the difference between the power correction amount and the photovoltaic power at the current moment as the active power reference value of the virtual synchronous generator for grid-type control of the inverter.

Images