JP4541828B2 - Control method of photovoltaic power generation system - Google Patents

Description

  The present invention relates to a photovoltaic power generation system that controls so that the output power value from a solar cell becomes a substantially maximum value, and in particular, a plurality of maximum values of output power that are generated due to uneven solar radiation on the panel surface of the solar cell. The present invention relates to a technique for searching for maximum power with high accuracy.

  FIG. 6 is a block diagram of a conventional photovoltaic power generation system. In the figure, SC1 to SC3 are solar cells, for example, three are connected in parallel. PT is an output voltage detection circuit that detects the output voltage of the solar cell, CT is an output current detection circuit that detects the output current of the solar cell, and CC is an MPPT controller that performs genetic algorithm control (GA) by a microprocessor. The microprocessor CC incorporates an AD converter for reading the output voltage value and output current value of the solar cell. IN is a DC / AC inverter circuit (or DC / DC inverter circuit) that converts a direct current output from a solar cell into a voltage, AD is a load, and SP is a commercial power supply. Here, the microprocessor of the controller CC calculates the output power from the solar cell SC by multiplying the output voltage and output current of the solar cell SC, and stores the voltage and power value in the memory. Further, the controller CC can control the output voltage from the solar cell SC by controlling the output of the DC / AC inverter circuit IN via the pulse width control circuit PWM.

  Next, genetic algorithm control (GA) will be described. The controller CC is a GA controller, regards a solar cell output set value (not shown) as a gene, stores an output power value from the solar cell as a gene evaluation value, and sets a plurality of values from a set range of the solar cell output set value. Genes are generated randomly or based on predetermined conditions to form a gene population, and the genes are selected and crossed and mutated to generate a predetermined number of genes to form a second generation population, By repeating the genetic algorithm operation, the next generation population is formed one after another, and the maximum power value is followed based on the value of the gene population.

  FIG. 7 is a flowchart for explaining the operation of the prior art. The operation will be described with reference to this flowchart and FIG. 8 in which the genes of the population converge by the genetic algorithm and follow the maximum value.

  A plurality of (for example, 10) genes are extracted at random or based on predetermined conditions from the set range of the solar cell output set value (step T1). Then, the counter i is set to 1 (step T2).

  A solar cell output set value corresponding to a predetermined gene among the plurality of genes is set as a first solar cell output set value, and an inverter is operated according to the first solar cell output set value to thereby output a first solar cell output. The voltage V1 is set (step T3). The first solar cell output current I1 output in accordance with the first solar cell output set value is measured (step T4). The microprocessor of the controller CC calculates the first solar cell output power W1 by multiplying the first solar cell output voltage V1 and the first solar cell output current I1, and stores it in the memory (step T5). . Then, 1 is added to the counter i. (Step T6)

  Next, it is determined whether all n (for example, 10) output powers of the solar cell output set values have been measured. If No, the process returns to step T3, and a predetermined number of the remaining plurality of genes is determined. The solar cell output set value corresponding to the gene is set as the second solar cell output set value, and the inverter is operated according to the second solar cell output set value to set the second solar cell output voltage V2 (step T3). ). The second solar cell output current I2 output according to the second solar cell output set value is measured (step T4). The microprocessor of the controller CC calculates the second solar cell output power W2 by multiplying the second solar cell output voltage V2 and the second solar cell output current I2, and stores it in the memory (step T4). Subsequently, 1 is added to the counter i (step T6).

  Thereafter, the same operation is repeated, the n-th solar cell output current In corresponding to the n-th solar cell output set value is measured, the n-th solar cell output power Wn is calculated and stored in the memory, and FIG. An evaluation value for the first generation initial population shown in A) is determined (step T7).

  Next, in the initial population, two genes are randomly extracted, the one having the larger evaluation value (solar cell output power) is selected, and the above selection is repeated to select two genes (step T8).

  The two selected genes are converted into a binary character string and crossed at an arbitrary point having a predetermined probability to generate two new genes (step T9).

  The crossed gene is mutated by intentionally changing a part of the character string with a predetermined low probability (step T10).

  When the number of selections and crossovers is n / 2 (for example, 10/2) or less, the process returns to step T8 (step T11). Then, from the initial population, two genes are randomly extracted again and the one with the larger evaluation value (solar cell output power) is selected, and the above selection is repeated to newly select two genes (step) T8).

  The two selected genes are converted into a binary character string and crossed at an arbitrary point having a predetermined probability to generate two new genes (step T9).

  The crossed gene is mutated by intentionally changing a part of the character string with a predetermined low probability (step T10).

  When the number of selections, crossovers, and mutations is n / 2 (for example, 10/2), the process proceeds to step 12, and a second generation population shown in FIG. 8B is formed by the generated genes.

  Thereafter, the generation of the gene population is renewed one after another by repeating the above operation, and the nth generation population is formed in FIG. 8C. The genes of the nth generation population are converged to the vicinity of the maximum value. Yes.

  When a plurality of maximum values are tracked using the genetic algorithm described above, it is possible to avoid the maximum value of a small mountain that is not the maximum power and to follow the maximum value of a large mountain. For example, Patent Document 1 is a prior document of genetic algorithm control disclosing the above-described technique.

Japanese Patent Application No. 2004-240257

  In the genetic algorithm of the prior art, even if a part of the solar cell panel is clouded and shaded, the solar radiation state changes, and multiple maximum values occur in the output power of the solar cell, It is possible to follow the maximum power of a plurality of maximum values.

  However, a large fluctuation in which a cloud is formed on a part of the panel of the solar cell to generate a shadow is several times a day. However, fluctuations in solar radiation or environmental temperature due to the movement of the sun always occur. When a shadow occurs in the solar cell, the maximum value (peak shape) of the output power fluctuates. However, since the number of fluctuations is small, the maximum value can be obtained by setting the mutation probability of the genetic algorithm to the optimum value. The latest maximum power can always be tracked according to the shape change of the value, but relatively small output fluctuations caused by fluctuations in solar radiation due to movement of the sun or relatively small output fluctuations caused by fluctuations in environmental temperature always occur. Also, since the speed of output fluctuation is relatively fast, the genetic algorithm with the relatively slow follow-up speed cannot sufficiently follow up.

  In order to improve the follow-up speed, it is conceivable to increase the probability of mutation to increase the follow-up speed. However, increasing the mutation probability increases the number of times the solar cell output set value fluctuates. For this reason, the fluctuation of the output power of the inverter device also becomes large.

  Then, this invention is providing the control method of the solar power generation system which can solve the said subject.

  In order to solve the above-described problem, the first invention controls the inverter so that the output voltage or output current from the solar cell is substantially equal to a predetermined solar cell output set value, and follows the change in the solar radiation state. And controlling the solar cell output set value to an appropriate value so that the output power value from the solar cell becomes a substantially maximum value, wherein the solar cell output set value is regarded as a gene. And a GA controller based on a genetic algorithm using the output power value from the solar cell as an evaluation value of the gene, and a plurality of genes are set randomly or in a predetermined condition from the set range of the solar cell output set value. Based on the extraction, a first generation initial population is formed, and the inverter is sequentially operated according to each solar cell output set value corresponding to each gene of the initial population, and the active solar cell is operated. Is stored as an evaluation value of each gene, and then each gene of the initial population is input to the GA controller, selected according to the evaluation value of each gene, and crossed and mutated to give a predetermined number of genes To generate a second generation population, and then sequentially operate the inverter according to each solar cell output set value corresponding to each gene of the second generation population and output power value from the operating solar cell Is stored as an evaluation value of each gene, and then each gene of the second generation population is input to the GA controller, selected according to the evaluation value of each gene, and crossover / mutated to output a predetermined number of genes Then, a third generation population is formed, and thereafter, by repeating the above operation, the generation of the gene population is renewed one after another by a genetic algorithm process for forming the nth generation population. In the control method of the photovoltaic power generation system that controls the output power value to be substantially the maximum value, before each gene of the n generation population is input to the GA controller, an evaluation value is determined in the n generation population. The largest gene is selected, and the solar cell output set value corresponding to the selected gene is set as the reference solar cell output set value, and the reference solar cell output set value and a value larger by a predetermined amount than the reference solar cell output set value Three solar cell output set values having a small fixed value are generated, the inverters are sequentially operated according to the respective solar cell output set values, and the magnitudes of the respective output power values are compared, and the maximum output power value is obtained. A solar cell output set value is selected, and then the solar cell output set value is set as the next reference solar cell output set value, and three solar cell output set values are set based on the reference solar cell output set value. The reference solar cell output set value is updated one after another by repeating the above operation, and when the new reference solar cell output set value becomes the same value continuously for a predetermined number of times, or the above-mentioned repeat count is preset. Sunlight characterized by adding a hill-climbing process for exchanging a gene corresponding to the reference solar cell output set value when a predetermined reference number of times is reached with a gene having the largest evaluation value in the n generation population It is a control method of a power generation system.

  According to a second aspect of the present invention, before operating the inverter with each solar cell output setting value corresponding to each gene of the n + 1 generation population, one gene of the n + 1 generation population is randomly extracted, The method for controlling a photovoltaic power generation system according to claim 1, wherein the gene is replaced with or added to a gene generated by a hill climbing process.

  According to a third aspect of the present invention, there is provided the solar power generation system control method according to claim 1, wherein the hill climbing process is performed when the n generation group becomes a generation group for each predetermined interval.

  According to the first invention, the solar cell shown in FIG. 5 (B) with respect to the relatively small output power fluctuation shown in FIG. 5 (A) caused by fluctuations in solar radiation amount due to movement of the sun or fluctuations in environmental temperature. The output set value is set, the inverter is operated sequentially to detect the largest output power value, and the vicinity of the maximum value of the output power value is followed by hill climbing processing with a high processing speed using the detected output power value as a reference value. . Subsequently, genetic algorithm processing is performed using a value in the vicinity of the maximum value, and each gene shown in FIG. 5C is converged to follow the maximum power value. As a result of a synergistic effect with the accuracy of following the maximum power value, it is possible to sufficiently follow the output power fluctuation with a relatively fast fluctuation speed caused by the fluctuation of the solar radiation amount or the fluctuation of the environmental temperature.

  According to the second invention, one gene of the n + 1 generation population formed by the genetic algorithm processing is randomly extracted and exchanged with the elite gene extracted by the n generation hill climbing processing step, When a new gene population is generated, improvement of the maximum power tracking accuracy can be expected due to the action of the exchanged elite gene.

  According to the third invention, the number of fluctuations of the solar cell output set value is greatly reduced by reducing the number of follow-ups of the maximum output power value by the hill-climbing process, so the inverter device of the first invention The fluctuation rate is smaller than the output power fluctuation.

[Embodiment 1]
FIG. 1 is a block diagram of a photovoltaic power generation system according to an embodiment of the present invention. In the figure, the same reference numerals as those in the block diagram of the conventional solar power generation system shown in FIG.

  In the block diagram of the photovoltaic power generation system shown in FIG. 1, a controller CC is a controller that performs genetic algorithm control (GA) and hill climbing control (HC), and is a microprocessor. Here, the microprocessor of the controller CC calculates the output power of the solar cell SC by multiplying the output voltage and the output current of the solar cell SC, and stores the output voltage and the output power value in the memory. In addition, the controller CC controls the output voltage from the solar cell SC by controlling the output of the DC / AC inverter IN via the pulse width control circuit PWM.

  Next, the GA / HC algorithm combining the genetic algorithm control (GA) and the hill climbing control (HC) of the controller CC will be described. First, the solar cell output set value (not shown) is regarded as a gene, the output power value from the solar cell is stored as a gene evaluation value, and a plurality of genes are randomly extracted from the set range of the solar cell output set value. To form the nth generation population. Next, the solar cell output set value corresponding to the gene having the largest evaluation value in the nth generation group is selected, and the process proceeds to the hill climbing process. The inverter is operated with the selected set value as the first reference solar cell output set value, and the output power value from the operating solar cell is calculated and stored, and then the first reference solar cell output setting is performed. A second solar cell output set value that is larger than the value by a predetermined amount is generated and an output power value is calculated and stored, and then a third solar cell output that is smaller than the first reference solar cell output set value by a predetermined amount. Generate a set value and calculate and store an output power value, compare the magnitude of each of the stored output power values, select a solar cell output set value corresponding to the largest output power value, The above solar cell output set value is set as the next reference solar cell output set value, and thereafter the above-described hill-climbing operation is repeated to obtain the reference solar cell output set value at which the output power value from the solar cell becomes the largest. That. Next, the gene corresponding to the obtained reference solar cell output power value is exchanged with the gene having the largest evaluation value in the nth generation population, and the gene evaluation value of the exchanged nth generation population is used. Performing genetic algorithm processing, selecting, crossing and mutating to generate a predetermined number of genes to form a new population, and repeating the above operation to update the generation of the gene population one after another and maximize the power value Follow.

  FIG. 2 is a flowchart for explaining the operation of the first embodiment of the present invention. The operation will be described with reference to this flowchart and the figure following the maximum value in FIG.

  From the set range of the solar cell output set value, n (for example, 10) genes are randomly extracted (step T1). Then, 1 is added to the counter i (step T2).

  Among the n genes, a solar cell output set value corresponding to a predetermined gene is set as a first solar cell output set value, and an inverter is operated according to the first solar cell output set value, thereby the first solar cell. The output voltage V1 is measured (step T3). The first solar cell output current I1 output in accordance with the first solar cell output set value is measured (step T4). The microprocessor of the controller CC calculates the first solar cell output power W1 by multiplying the first solar cell output voltage V1 and the first solar cell output current I1, and stores it in the memory (step T5). . Then, 1 is added to the counter i. (Step T6).

  Next, it is determined whether all n (for example, 10) output power values of the solar cell output have been measured. If No, the process returns to step T3, and among the remaining n-1 genes. The solar cell output set value corresponding to the predetermined gene is set as the second solar cell output set value, and the inverter is operated according to the second solar cell output set value to measure the second solar cell output voltage V2 (step) T3). The second solar cell output current I2 output according to the second solar cell output set value is measured (step T4). The microprocessor of the controller CC calculates the second solar cell output power W2 by multiplying the second solar cell output voltage V2 and the second solar cell output current I2, and stores it in the memory (step T4). Subsequently, 1 is added to the counter i (step T6).

  Thereafter, the same operation is repeated, and when all n (for example, 10) output power values of the solar cell output are measured, the result shown in FIG. Form an initial population of the first generation shown. (Step T7).

  Among the stored solar cell output power values, the gene corresponding to the maximum value is selected and the process jumps to Step T15 (Step T8). The solar cell output set value corresponding to the selected gene is set as the first reference solar cell output set value (step T15).

  The inverter is operated in accordance with the first reference solar cell output set value, and the first reference solar cell output voltage V1 shown in FIG. 4B is measured (step T16). The first reference solar cell output current I1 output in accordance with the first reference solar cell output set value is measured (step T17). The microprocessor of the controller CC calculates the first reference solar cell output power W1 by multiplying the first reference solar cell output voltage V1 and the first reference solar cell output current I1, and stores the first reference solar cell output power W1 in the memory. The reference solar cell output voltage V1 and the first reference solar cell output power W1 are stored (step T18).

  Next, when a second solar cell output set value that is a predetermined amount larger than the first reference solar cell output set value is set and the inverter is operated, the first reference solar cell output voltage shown in FIG. V1 becomes the second solar cell output voltage V2 (step T19).

  The second solar cell output voltage V2 and the second solar cell output current I2 output in accordance with the second solar cell output set value are measured (step T20). The microprocessor of the controller CC calculates the second solar cell output power W2 by multiplying the second solar cell output voltage V2 and the second solar cell output current I2, and stores the second solar cell in the memory. The values of the output voltage V2 and the second solar cell output power W2 are stored (step T21).

  Subsequently, when a third solar cell output set value that is a predetermined amount smaller than the first reference solar cell output set value is set and the inverter is operated, the first reference solar cell output voltage V1 shown in FIG. Becomes the third solar cell output voltage V3 (step T22).

  The third solar cell output voltage V3 and the third solar cell output current I3 output in accordance with the third solar cell output set value are measured (step T23). The microprocessor of the controller CC calculates the third solar cell output power W3 by multiplying the third solar cell output voltage V3 and the third solar cell output current I3, and stores the third solar cell in the memory. The values of the output voltage V3 and the third solar cell output power W3 are stored (step T24).

  The first reference solar cell output voltage V1, the second solar cell output voltage V2, and the third solar cell output voltage V3 shown in FIG. 4B stored as described above are output at (V3 <V1 <V2). The power values W1, W2, and W3 are compared with each other, the solar cell output set value corresponding to the largest output power value is selected, and set as the next reference solar cell output set value (step T25).

  It is determined whether the reference solar cell output set value is the same value continuously for a predetermined number of times, or whether the reference number of times is repeated a predetermined number of times by repeating the operation from step T16. If NO, the process returns to step T16. . Subsequently, it is determined whether the reference solar cell output setting value obtained by repeating the operation from step T16 to step T25 is the same value continuously for a predetermined number of times or a predetermined reference number, and if Yes, Proceed to step T9 (step T67).

  Returning to the genetic algorithm process from the hill-climbing process, first, the gene according to the reference solar cell output setting value selected by the hill-climbing process and the gene having the largest evaluation value in the initial population, n new gene populations are generated (step T9).

  Next, the n new gene populations are input to the GA controller, and two genes out of the n genes are randomly extracted and the solar cell output power value that is the evaluation value is larger. Select and repeat the above selection to select two genes (step T10).

  The two selected genes are converted into a binary character string and crossed at an arbitrary point having a predetermined probability to generate two new genes (step T11).

  The crossed gene is mutated by intentionally changing a part of the character string with a predetermined low probability (step T12).

  When the number of selections and crossovers is n / 2 (for example, 10/2) or less, the process returns to step T10 (step T13). When the number of times of selection and crossover is n / 2 (for example, 10/2), the process proceeds to step 14 to form the second generation group shown in FIG. 4C (step T14).

  Subsequently, returning to step T2, the operation from step T3 to step T7 described above is repeated, and the inverter is sequentially operated according to each solar cell output set value corresponding to each gene of the second generation population, and the solar cell in operation Is stored as an evaluation value for each gene. In step T8, the gene corresponding to the maximum value is selected from the stored solar cell output powers, and the process jumps to step T15 (step T8). The solar cell output set value corresponding to the selected gene is set as the first reference solar cell output set value (step T15). Thereafter, by repeating the same hill climbing process and genetic algorithm process as described above, the generation of the gene population is renewed one after another to form the nth generation population shown in FIG. 4 (D), and the output power from the solar cell The value is controlled so as to be approximately the maximum value.

  Using the GA / HC algorithm that combines the genetic algorithm control (GA) and hill climbing control (HC) described above, it is possible to perform control that combines the speed of hill-climbing processing with good tracking accuracy of genetic algorithm processing. Become.

[Embodiment 2]
The operation of the second embodiment will be described using the flowchart shown in FIG. In this figure, the same reference numerals as those in the flowchart of the first embodiment of the present invention shown in FIG. Further, the only difference between the present invention and the second embodiment is step T14 in the flowchart shown in FIG.

  In step T26, the process returns from the hill-climbing process to the genetic algorithm process. First, the gene corresponding to the reference solar cell output power value set by the hill-climbing process is replaced with the gene having the largest evaluation value in the initial population. N gene populations are generated (step T9).

  Next, the n new gene populations are input to the GA controller, and two genes out of the n genes are randomly extracted and the solar cell output power value that is the evaluation value is larger. Select and repeat the above selection to select two genes (step T10).

  The two selected genes are converted into a binary character string and crossed at an arbitrary point having a predetermined probability to generate two new genes (step T11).

  The crossed gene is mutated by intentionally changing a part of the character string with a predetermined low probability (step T12).

  When the number of selections and crossovers is n / 2 (for example, 10/2) or less, the process returns to step T10 (step T13). When the number of selections and crossovers is n / 2 (for example, 10/2), the process proceeds to step 14 (step T13).

  Subsequently, one gene of the n + 1 generation population formed by the genetic algorithm processing is randomly extracted and replaced or added with the gene generated by the n generation hill climbing processing to generate an n + 1 generation population. Then, returning to step T2, the same operation as in the first embodiment is performed (step 14).

[Embodiment 3]
The operation of the third embodiment will be described using the flowchart shown in FIG. In this figure, the same reference numerals as those in the flowchart of the first embodiment of the present invention shown in FIG. Further, the difference between the present invention and the third embodiment is only step T8 in the flowchart of FIG.

  In step T7, when all n (for example, 10) output power values of the solar cell output values are measured, the first solar cell output power value shown in FIG. An initial population of generations is formed (step T7).

  When the number of genetic algorithm processes is counted and the number of genetic algorithm processes reaches a predetermined number of predetermined intervals, the maximum value is selected from the solar cell output power values stored in step T7, and the process jumps to step T15. Then, the mountain climbing process is performed. If the number of genetic algorithm steps is not a predetermined number of predetermined intervals, the process proceeds to step T9. Subsequently, the determined n gene groups are input to the GA controller to perform genetic algorithm processing. (Step T9).

1 is a block diagram of a photovoltaic power generation system according to an embodiment of the present invention. It is a flowchart explaining the operation | movement of Embodiment 1 of this invention. It is a flowchart explaining operation | movement of Embodiment 2, 3 of this invention. FIG. 4 is a diagram for tracking a maximum value according to the first embodiment. FIG. 10 is a diagram for tracking a maximum value according to the second embodiment. It is a block diagram of the solar power generation system of a prior art. It is a flowchart explaining operation | movement of a prior art. It is a figure which tracks maximum value by a prior art.

Explanation of symbols

AD load CC controller CT current detection circuit IN inverter circuit PT voltage detection circuit PWM pulse width control circuit SP system power supply SC1 solar cell SC2 solar cell SC3 solar cell

Claims (3)

  The inverter is controlled so that the output voltage or output current from the solar cell becomes substantially equal to a predetermined solar cell output set value, and the output power value from the solar cell becomes a substantially maximum value following the change in the solar radiation state. A control method for a photovoltaic power generation system that controls the solar cell output set value to an appropriate value, wherein the solar cell output set value is regarded as a gene and the output power value from the solar cell is used as a gene evaluation value A GA controller based on a genetic algorithm is provided, and a plurality of genes are extracted from a set range of the solar cell output set value randomly or based on a predetermined condition to form a first generation initial population, The inverter is sequentially operated according to each solar cell output set value corresponding to each gene of this initial population, and the output power value from the operating solar cell is stored as an evaluation value of each gene. Subsequently, each gene of the initial population is input to the GA controller, selected according to the evaluation value of each gene, and crossed and mutated to output a predetermined number of genes to form a second generation population, , Sequentially operating the inverter with each solar cell output set value corresponding to each gene of the second generation population and storing the output power value from the operating solar cell as an evaluation value of each gene, Each gene of the second generation population is input to the GA controller, selected according to the evaluation value of each gene, crossed and mutated, and a predetermined number of genes are output to form a third generation population. The generation of the gene population is renewed one after another by repeating the above, and the output power value from the solar cell is controlled to a substantially maximum value by the genetic algorithm process of forming the nth generation population. In the control method of the photovoltaic system, before inputting each gene of the n generation population to the GA controller, the gene having the largest evaluation value is selected from the n generation population, and the gene corresponding to the selected gene is selected. The solar cell output set value is set as a reference solar cell output set value, and three solar cell output set values, the reference solar cell output set value, a value larger by a predetermined amount than the reference solar cell output set value, and a value smaller by a predetermined amount are generated. The inverters are sequentially operated according to the respective solar cell output set values to compare the magnitudes of the respective output power values, and the solar cell output set value that becomes the maximum output power value is selected. The battery output set value is set as the next reference solar cell output set value, three solar cell output set values are generated based on the reference solar cell output set value, and the above operations are repeated to successively change the basic output. The reference when the quasi-solar cell output set value is renewed and the new reference solar cell output set value becomes the same value continuously for a predetermined number of times, or when the number of repetitions reaches a predetermined reference number A control method for a photovoltaic power generation system, characterized in that a hill-climbing process step for exchanging a gene corresponding to a solar cell output set value with a gene having the largest evaluation value in the n generation population is added.   Before operating the inverter according to each solar cell output set value corresponding to each gene of the n + 1 generation population, one gene of the n + 1 generation population is randomly extracted and generated by the hill climbing process step of the n generation. The method for controlling a solar power generation system according to claim 1, wherein the method is replaced with or added to a gene. The method of controlling a photovoltaic power generation system according to claim 1, wherein the hill-climbing process step is performed when the n generation group becomes a generation group for every predetermined interval.

Images