Determination device, determination method, and determination program

Abstract

To further accurately detect an anomaly of a power generation unit in a photovoltaic power generation system.SOLUTION: A determination device comprises: a threshold acquisition unit for acquiring a threshold level to be set based on the amount of power generation in a target part of a solar cell panel; an actual value acquisition unit for acquiring an actual value of the output of each string; an estimated value calculation unit for calculating an estimated value of the output of the target string at a target time on the basis of the actual value of the output of the target string and the actual value of the output of a reference string; a feature quantity calculation unit for calculating a feature quantity that is a ratio between the estimated value and the actual value of the output of the target string at the target time; and a determination unit for performing determination processing to determine an output drop equivalent to the amount of power generation of the target string in the target part on the basis of the comparison result between the threshold level and the figure of merit based on the feature quantity.SELECTED DRAWING: Figure 8

Description

[Technical field]

[0001] The present disclosure relates to a determination device, a determination method, and a determination program. [Background technology]

[0002] In recent years, against the background of the aging of photovoltaic power generation systems, an increase in the number of photovoltaic power generation systems, and a shortage of managers of photovoltaic power generation systems, a technology for automatically monitoring the output of a photovoltaic power generation system has been developed. For example, Patent Document 1 (JP Patent Publication No. 2012-205078) discloses a photovoltaic power generation monitoring system as follows. That is, the photovoltaic power generation monitoring system is a photovoltaic power generation monitoring system for monitoring the power generation status of a plurality of photovoltaic panels in a photovoltaic power generation system that aggregates the output from the plurality of photovoltaic panels and sends it to a power conversion device, and includes a measuring device that is provided at a location where the output electric paths from the plurality of photovoltaic panels are aggregated and measures the amount of power generated by each photovoltaic panel, a lower-side communication device that is connected to the measuring device and has a function of transmitting measurement data of the amount of power generated by the measuring device, an upper-side communication device that has a function of receiving the measurement data transmitted from the lower-side communication device, and a management device that has a function of collecting the measurement data for each photovoltaic panel via the upper-side communication device.

[0003] Furthermore, Patent Document 2 (International Publication No. 2019 / 187525) discloses the following determination device. That is, the determination device is used in a solar power generation system equipped with a plurality of power generation units including a solar panel, and includes an acquisition unit that acquires a plurality of pieces of measurement information indicating the measurement results of the output current of the plurality of power generation units, a detection unit that detects the occurrence of a state caused by overloading of the solar power generation system, and a determination unit that determines an abnormality in the power generation unit corresponding to the output current based on a fluctuation in the output current when the state occurs. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2012-205078 A [Patent Document 2] International Publication No. 2019 / 187525 Summary of the Invention [Problem to be solved by the invention]

[0005] There is a need for a technique that goes beyond the techniques described in Patent Documents 1 and 2 and is capable of more accurately detecting an abnormality in a power generation unit in a photovoltaic power generation system.

[0006] The present disclosure has been made to solve the above-mentioned problems, and its purpose is to provide a determination device, a determination method, and a determination program that are capable of more accurately detecting abnormalities in a power generation unit in a solar power generation system. [Means for solving the problem]

[0007] The determination device disclosed herein is a determination device used in a solar power generation system having a plurality of strings including solar panels, and includes a threshold acquisition unit that acquires a threshold set based on the magnitude of the output of a target portion of the solar panel, an actual value acquisition unit that acquires an actual value of the output of each of the strings, a predicted value calculation unit that calculates a predicted value of the output of the target string at a target time based on the actual value of the output of a target string among the plurality of strings and the actual value of the output of a reference string among the plurality of strings acquired by the actual value acquisition unit, a feature calculation unit that calculates a feature that is the ratio between the predicted value calculated by the predicted value calculation unit and the actual value of the output of the target string at the target time, and a determination unit that performs a determination process to determine an output reduction in the target string that corresponds to the amount of power generation of the target portion based on a comparison result between an evaluation value based on the feature calculated by the feature calculation unit and the threshold acquired by the threshold acquisition unit.

[0008] One aspect of the present disclosure can be realized not only as a determination device having such a characteristic processing unit, but also as a semiconductor integrated circuit that realizes part or all of the determination device, or as a system that includes the determination device. Effect of the Invention

[0009] According to the present disclosure, an abnormality in a power generation unit in a solar power generation system can be detected more accurately. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram showing a configuration of a solar power generation system according to an embodiment of the present disclosure. [Diagram 2] FIG. 2 is a diagram illustrating a configuration of a PCS unit according to an embodiment of the present disclosure. [Diagram 3] FIG. 3 is a diagram illustrating a configuration of a current collecting unit according to an embodiment of the present disclosure. [Figure 4] FIG. 4 is a diagram illustrating a configuration of a solar cell unit according to an embodiment of the present disclosure. [Diagram 5] FIG. 5 is a diagram illustrating a configuration of a solar cell panel according to an embodiment of the present disclosure. [Figure 6] FIG. 6 is a diagram illustrating a configuration of a power generation state management system according to an embodiment of the present disclosure. [Figure 7] FIG. 7 is a diagram illustrating a configuration of a monitoring device in a power generation state management system according to an embodiment of the present disclosure. [Figure 8] FIG. 8 is a diagram illustrating a configuration of a detection device according to an embodiment of the present disclosure. [Figure 9] FIG. 9 is a diagram illustrating an example of a method for setting the lower threshold value ThL stored in the storage unit in the detection device according to the embodiment of the present disclosure. [Figure 10] FIG. 10 is a diagram illustrating an example of a method for setting the upper limit threshold ThH stored in the storage unit in the detection device according to the embodiment of the present disclosure. [Figure 11]FIG. 11 is a flowchart defining an example of an operation procedure when the detection device according to the embodiment of the present disclosure performs the detection process. [Figure 12] FIG. 12 is a flowchart defining an example of an operation procedure when the detection device according to the embodiment of the present disclosure performs a main process. [Figure 13] FIG. 13 is a flowchart defining an example of an operation procedure when the detection device according to the embodiment of the present disclosure performs post-processing. [Figure 14] FIG. 14 is a flowchart defining an example of an operation procedure when the detection device according to the embodiment of the present disclosure performs post-processing. [Figure 15] FIG. 15 is a flowchart illustrating an example of an operation procedure when a detection device according to an embodiment of the present disclosure performs post-processing. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] First, the contents of the embodiments of the present disclosure will be listed and described.

[0012] (1) A determination device according to an embodiment of the present disclosure is a determination device used in a solar power generation system having a plurality of strings including solar panels, and includes: a threshold acquisition unit that acquires a threshold that is set based on the magnitude of the output of a target portion of the solar panel; an actual value acquisition unit that acquires an actual value of the output of each of the strings; a predicted value calculation unit that calculates a predicted value of the output of the target string at a target time based on the actual value of the output of a target string among the plurality of strings and the actual value of the output of a reference string among the plurality of strings acquired by the actual value acquisition unit; a feature calculation unit that calculates a feature that is a ratio between the predicted value calculated by the predicted value calculation unit and the actual value of the output of the target string at the target time; and a determination unit that performs a determination process to determine an output reduction in the target string that corresponds to the amount of power generated by the target portion, based on a result of comparing an evaluation value based on the feature calculated by the feature calculation unit with the threshold acquired by the threshold acquisition unit.

[0013] In this way, by using a configuration that determines the output drop in the target string based on the comparison result between an evaluation value based on the predicted and actual values ​​of the output of the target string and a threshold value that is set based on the amount of power generation in the target portion of the solar panel, it is possible to more accurately determine the output drop corresponding to the amount of power generation in the target portion even if the rate of drop in the output of the target string when power generation in the target portion stops is small. This makes it possible to accurately detect an abnormality in the target portion of the target string that is determined to have a reduced output. Therefore, it is possible to more accurately detect an abnormality in the power generation unit in the solar power generation system.

[0014] (2) In the above (1), the target part may include a bypass circuit, and the threshold value may be set based on the output characteristics of a normal string, which is the string including the solar panel in which the output of the target part is normal, and the output characteristics of a reduced output string, which is the string including the solar panel in which the output of the target part has decreased due to a short circuit in the bypass circuit.

[0015] With this configuration, a drop in output from the target portion due to a short circuit in the bypass circuit can be detected more accurately.

[0016] (3) In (2) above, the threshold may be set based on the output of the normal string at a third voltage value and the output of the reduced output string at the third voltage value, and the third voltage value may be greater than a first voltage value corresponding to a peak power in a power-voltage characteristic of the normal string and less than a value obtained by adding an absolute value of the difference between the first voltage value and a second voltage value corresponding to a peak power in a power-voltage characteristic of the reduced output string to the first voltage value.

[0017] With this configuration, it is possible to more accurately determine the output reduction of the target string due to a short circuit in the bypass circuit, based on the change in the power-voltage characteristics due to the short circuit in the bypass circuit.

[0018] (4) In any of (1) to (3) above, the predicted value calculation unit may calculate the predicted value by excluding the actual values ​​for a time period in which at least one of the actual value of the output of the target string, the actual value of the output of the reference string, and the ratio between the actual value of the output of the target string and the actual value of the output of the reference string satisfies a predetermined condition.

[0019] This configuration allows, for example, the actual output of each string to be excluded during periods of extreme weather and localized shadowing, allowing a more accurate prediction of the output of the target string to be calculated.

[0020] (5) In any of (1) to (4) above, the judgment unit may perform the judgment process using the evaluation value based on the feature at a plurality of target times in a target period, and the prediction value calculation unit may calculate the prediction value based on the actual value prior to the target period.

[0021] With this configuration, a more accurate predicted value of the output of the target string can be calculated based on actual values ​​from a period in the past, prior to the target judgment period in which an abnormality may have already occurred in the target string, when there is a lower possibility that an abnormality has occurred in the target string.

[0022] (6) In any of (1) to (5) above, the judgment unit may determine whether to validate a judgment result that the output of the target string has decreased in the judgment process, based on an output ratio that is a ratio between a representative value of the actual values ​​of the outputs of the multiple strings and the actual value of the output of the target string.

[0023] With this configuration, the appropriateness of the judgment result in the judgment process can be determined based on the relationship between the actual value of the output of the target string and the representative value of the actual values ​​of the outputs of the other strings, and an erroneous judgment result can be invalidated.

[0024] (7) In the above (6), the judgment unit may invalidate the judgment result when a difference between a plurality of the output ratios corresponding to a plurality of measurement timings of the actual value and an output ratio within a predetermined range among the plurality of output ratios is equal to or greater than a predetermined value.

[0025] With this configuration, when it is determined that an output reduction corresponding to the amount of power generated by the target part has occurred in the target string, it can be determined that this is due to the influence of shading, and the determination result in the determination process can be invalidated.

[0026] (8) In the above (6) or (7), the determination unit may invalidate the determination result when a rate of decrease in the output ratio in a period before the target time is equal to or greater than a predetermined value.

[0027] With this configuration, when a decrease in output of a target string that has been determined to have experienced an output decrease equivalent to the power generation amount of the target part has been progressing before the judgment timing, it is possible to determine that the output decrease is due to the influence of shading and invalidate the judgment result in the judgment process.

[0028] (9) In any one of (6) to (8) above, the determination unit may invalidate the determination result when a rate of decrease in the output ratio in a period after the target time is equal to or greater than a predetermined value.

[0029] With this configuration, if the output reduction of the target string that has been determined to have experienced an output reduction equivalent to the power generation amount of the target part progresses after the judgment timing, it is possible to determine that the output reduction is due to the influence of shading, and invalidate the judgment result in the judgment process.

[0030] (10) In any of (6) to (9) above, the judgment unit may invalidate the judgment result when an increase rate of the output ratio within a predetermined range among a plurality of the output ratios corresponding to a plurality of measurement timings in a period after the target time is equal to or greater than a predetermined value.

[0031] With this configuration, if the output of the target string that has been determined to have experienced an output decrease equivalent to the power generation amount of the target part increases after the judgment timing, it is possible to determine that the malfunction of the target part has been resolved and invalidate the judgment result in the judgment process.

[0032] (11) In any of (1) to (10) above, the judgment unit may determine whether to validate a judgment result that the output of the target string has decreased in the judgment process, based on the length of a period during which a proportion of the actual values ​​indicating an output decrease corresponding to the power generation amount of the target part in the target string, among the multiple actual values ​​in a period including the target time, continues to be equal to or greater than a predetermined value.

[0033] With this configuration, the appropriateness of the judgment result in the judgment process can be judged based on the change in the actual value of the output of the target string before and after the judgment timing at which it is determined that an output decrease equivalent to the power generation amount of the target part has occurred, and an erroneous judgment result can be invalidated.

[0034] (12) A determination method according to an embodiment of the present disclosure is a determination method in a determination device used in a solar power generation system having a plurality of strings including solar panels, and includes the steps of: acquiring a threshold value that is set based on the magnitude of the output of a target portion of the solar panel; acquiring actual values ​​of the output of each of the strings; calculating a predicted value of the output of the target string at a target time based on the acquired actual value of the output of a target string among the plurality of strings and the acquired actual value of the output of a reference string among the plurality of strings; calculating a feature that is a ratio between the calculated predicted value and the actual value of the output of the target string at the target time; and performing a determination process to determine an output reduction in the target string that corresponds to the amount of power generated by the target portion based on a result of comparing an evaluation value based on the calculated feature with the acquired threshold.

[0035] In this way, by using a method for determining an output drop in a target string based on a comparison between an evaluation value based on the predicted and actual values ​​of the output of the target string and a threshold value set based on the magnitude of the power generation amount of the target portion of the solar panel, it is possible to more accurately determine an output drop corresponding to the power generation amount of the target portion even if the rate of drop in the output of the target string when power generation in the target portion stops is small. This makes it possible to accurately detect an abnormality in the target portion of the target string determined to have a reduced output. Therefore, it is possible to more accurately detect an abnormality in the power generation unit in the solar power generation system.

[0036] (13) A judgment program according to an embodiment of the present disclosure is a judgment program used in a judgment device used in a solar power generation system having a plurality of strings including solar panels, and causes a computer to function as: a threshold acquisition unit that acquires a threshold set based on the magnitude of the output of a target portion of the solar panel; an actual value acquisition unit that acquires an actual value of the output of each of the strings; a predicted value calculation unit that calculates a predicted value of the output of the target string at a target time based on the actual value of the output of a target string among the plurality of strings and the actual value of the output of a reference string among the plurality of strings acquired by the actual value acquisition unit; a feature calculation unit that calculates a feature that is the ratio between the predicted value calculated by the predicted value calculation unit and the actual value of the output of the target string at the target time; and a judgment unit that performs a judgment process to determine an output reduction in the target string corresponding to the amount of power generation of the target portion based on a comparison result between an evaluation value based on the feature calculated by the feature calculation unit and the threshold acquired by the threshold acquisition unit.

[0037] In this way, by using a configuration that determines the output drop in the target string based on the comparison result between an evaluation value based on the predicted and actual values ​​of the output of the target string and a threshold value that is set based on the amount of power generation in the target portion of the solar panel, it is possible to more accurately determine the output drop corresponding to the amount of power generation in the target portion even if the rate of drop in the output of the target string when power generation in the target portion stops is small. This makes it possible to accurately detect an abnormality in the target portion of the target string that is determined to have a reduced output. Therefore, it is possible to more accurately detect an abnormality in the power generation unit in the solar power generation system.

[0038] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. In the drawings, the same or corresponding parts are denoted by the same reference numerals, and the description thereof will not be repeated. In addition, at least some of the embodiments described below may be arbitrarily combined.

[0039] <Configuration and basic operation> [Solar power generation system configuration] FIG. 1 is a diagram showing a configuration of a solar power generation system according to an embodiment of the present disclosure.

[0040] 1, a photovoltaic power generation system 401 includes four PCS (Power Conditioning System) units 80 and a cubicle 6. The cubicle 6 includes a copper bar 73.

[0041] In FIG. 1, four PCS units 80 are representatively shown, but a greater or lesser number of PCS units 80 may be provided.

[0042] FIG. 2 is a diagram illustrating a configuration of a PCS unit according to an embodiment of the present disclosure.

[0043] 2, the PCS unit 80 includes four current collecting units 60 and a PCS 8. The PCS 8 includes a copper bar 7 and a power conversion unit 9.

[0044] In FIG. 2, four current collecting units 60 are representatively shown, but a greater or lesser number of current collecting units 60 may be provided.

[0045] FIG. 3 is a diagram illustrating a configuration of a current collecting unit according to an embodiment of the present disclosure.

[0046] 3, the current collecting unit 60 includes four solar cell units 75 and a current collecting box 71. The current collecting box 71 has a copper bar 72.

[0047] In FIG. 3, four solar cell units 75 are representatively shown, but a greater or lesser number of solar cell units 75 may be provided.

[0048] FIG. 4 is a diagram illustrating a configuration of a solar cell unit according to an embodiment of the present disclosure.

[0049] 4, the solar cell unit 75 includes four power generation sections 78 and a junction box 76. The power generation sections 78 have a solar cell panel 79. The junction box 76 has a copper bar 77.

[0050] In FIG. 4, four power generation units 78 are representatively shown, but a greater or lesser number of power generation units 78 may be provided.

[0051] In this example, the power generation unit 78 is a string in which 14 solar cell panels 79 are connected in series. Note that a larger or smaller number of solar cell panels 79 may be provided in one power generation unit 78.

[0052] In the photovoltaic power generation system 401, the output lines 1 and the aggregated lines 2 and 5, that is, the power lines, from the multiple power generation units 78 are electrically connected to the PCS 8 shown in FIG.

[0053] In PCS8, for example, when power conversion unit 9 receives DC power generated in each power generation unit 78 via output line 1, copper bar 77, aggregation line 5, copper bar 72, aggregation line 2, copper bar 7 and internal line 3, it converts the received DC power into AC power and outputs it to aggregation line 4.

[0054] The aggregated line 4 has a first end connected to the power conversion unit 9 and a second end connected to the copper bar 73 .

[0055] In the cubicle 6, the AC power output from the power conversion unit 9 in each PCS 8 to each aggregation line 4 is output to the system via the copper bar 73.

[0056] 5 is a diagram showing a configuration of a solar cell panel according to an embodiment of the present disclosure. The arrows in FIG. 5 indicate the direction of current flowing through a solar cell panel 79.

[0057] 5, the solar cell panel 79 includes clusters 92A, 92B, and 92C. The cluster 92A has a bypass diode 93A and a plurality of solar cells 91. The cluster 92B has a bypass diode 93B and a plurality of solar cells 91. The cluster 92C has a bypass diode 93C and a plurality of solar cells 91. Hereinafter, each of the clusters 92A, 92B, and 92C will also be referred to as a cluster 92, and each of the bypass diodes 93A, 93B, and 93C will also be referred to as a bypass diode 93. The cluster 92 is an example of a target portion. The bypass diode 93 is an example of a bypass circuit. The plurality of solar cells 91 in the solar cell panel 79 are connected in series.

[0058] An anode of the bypass diode 93C is connected to a node N6 between a first end of the series circuit of the solar cell 91 in the cluster 92C and the adjacent solar cell panel 79 on the upstream side in the current direction. A cathode of the bypass diode 93C is connected to a node N5 between the second end of the series circuit and the cluster 92B.

[0059] The anode of the bypass diode 93B is connected to a node N4 between a first end of the series circuit of the solar cell 91 in the cluster 92B and a node N5. The cathode of the bypass diode 93B is connected to a node N3 between a second end of the series circuit and the cluster 92A.

[0060] An anode of the bypass diode 93A is connected to a node N2 between a first end of the series circuit of the solar cell cells 91 in the cluster 92A and a node N3. A cathode of the bypass diode 93A is connected to a node N1 between a second end of the series circuit and the adjacent solar cell panel 79 on the downstream side in the current direction.

[0061] [Configuration of power generation status management system] FIG. 6 is a diagram illustrating a configuration of a power generation state management system according to an embodiment of the present disclosure.

[0062] 6, the power generation state management system 301 is used in a solar power generation system 401. The power generation state management system 301 includes a detection device 101, a plurality of monitoring devices 111, and a plurality of collection devices 151. The detection device 101 is an example of a determination device.

[0063] In Figure 6, four monitoring devices 111 are shown representatively corresponding to one current collecting unit 60, but a greater or lesser number of monitoring devices 111 may be provided corresponding to one current collecting unit 60.

[0064] In the power generation state management system 301, information from a sensor in the monitoring device 111, which is a slave device, is transmitted to the collecting device 151, which is a master device, on a regular or irregular basis.

[0065] The monitoring device 111 is provided, for example, in the power collecting unit 60. More specifically, four monitoring devices 111 are provided corresponding to the four solar cell units 75, respectively. Each monitoring device 111 is electrically connected to, for example, the corresponding output line 1 and aggregation line 5.

[0066] The monitoring device 111 monitors the power generation unit 78 in the corresponding solar cell unit 75. For example, the monitoring device 111 measures the current of each output line 1 in the corresponding solar cell unit 75 using a sensor. Also, for example, the monitoring device 111 measures the voltage of the output line 1 in the corresponding solar cell unit 75 using a sensor.

[0067] The collecting device 151 collects measurement information indicating the measurement results by the monitoring device 111. The collecting device 151 is provided, for example, near the PCS 8. More specifically, the collecting device 151 is provided corresponding to the PCS unit 80, and is electrically connected to the copper bar 7 in the PCS 8 via the conductor 46. The collecting device 151 collects measurement information from a plurality of monitoring devices 111 provided in the current collecting unit 60 in the corresponding PCS unit 80.

[0068] The monitoring device 111 and the collecting device 151 transmit and receive information by performing power line communication (PLC) via the aggregation lines 2 and 5.

[0069] More specifically, each monitoring device 111 measures the current and voltage of the corresponding output line 1 and generates measurement information including the measurement results. Each monitoring device 111 transmits the generated measurement information to the collecting device 151 via the aggregation lines 2 and 5.

[0070] The collection device 151 performs power line communication with the monitoring device 111 via the aggregation lines 2 and 5 and receives measurement information from the monitoring device 111 .

[0071] The collecting device 151 transmits the measurement information received from the monitoring device 111 to the detecting device 101, for example, via a Virtual Private Network (VPN).

[0072] [Monitoring device configuration] 7 is a diagram showing the configuration of a monitoring device in a power generation state management system according to an embodiment of the present disclosure, in which the inside of a junction box 76 is shown in more detail.

[0073] 7, each of the output lines 1 includes a positive output line 1p and a negative output line 1n. The aggregated line 5 includes a positive aggregated line 5p and a negative aggregated line 5n. The copper bar 77 includes a positive copper bar 77p and a negative copper bar 77n.

[0074] Although not shown, the copper bars 72 in the current collection box 71 shown in FIG. 3 include a positive copper bar 72p and a negative copper bar 72n corresponding to the positive aggregated line 5p and the negative aggregated line 5n, respectively.

[0075] The positive output line 1p has a first end connected to the corresponding power generating unit 78 and a second end connected to the positive copper bar 77p. The negative output line 1n has a first end connected to the corresponding power generating unit 78 and a second end connected to the negative copper bar 77n.

[0076] The positive-side aggregated line 5p has a first end connected to the positive-side copper bar 77p and a second end connected to the positive-side copper bar 72p in the current collection box 71. The negative-side aggregated line 5n has a first end connected to the negative-side copper bar 77n and a second end connected to the negative-side copper bar 72n in the current collection box 71.

[0077] The monitoring device 111 includes an acquisition unit 11, a communication unit 14, four current sensors 16, and a voltage sensor 17. Note that the monitoring device 111 may include a greater or lesser number of current sensors 16 depending on the number of output lines 1, and may include, for example, 20 current sensors 16.

[0078] The monitoring device 111 is provided, for example, near the power generation unit 78. Specifically, the monitoring device 111 is provided, for example, inside the connection box 76 in which the copper bar 77 to which the output line 1 to be measured is connected is provided. The monitoring device 111 may be provided outside the connection box 76.

[0079] The monitoring device 111 is electrically connected to the positive aggregated line 5p and the negative aggregated line 5n via a positive power line 26p and a negative power line 26n, respectively. Hereinafter, each of the positive power line 26p and the negative power line 26n will also be referred to as a power line 26.

[0080] The current sensor 16 is, for example, a Hall element type current probe connected between the power generation unit 78 and the PCS 8. The current sensor 16 measures the current C flowing through the corresponding negative output line 1n. The current sensor 16 outputs a signal at a level according to the current C to the acquisition unit 11. The current sensor 16 may also measure the current C flowing through the positive output line 1p.

[0081] The voltage sensor 17 measures the voltage V between the positive side copper bar 77p and the negative side copper bar 77n. The voltage sensor 17 outputs a signal at a level corresponding to the voltage V to the acquisition unit 11.

[0082] The acquisition unit 11 acquires the measurement result of the output of the power generation unit 78. More specifically, the acquisition unit 11 acquires the measurement result of the current C by the current sensor 16 and the measurement result of the voltage V by the voltage sensor 17 as the measurement result of the output of the power generation unit 78. The acquisition unit 11 acquires a current measurement value CMtm that is the measurement value of the current C by each current sensor 16 based on a signal output from each current sensor 16 during a measurement period Pm from 8:00 to 16:00, for example. The acquisition unit 11 also acquires a voltage measurement value VMtm that is the measurement value of the voltage V by the voltage sensor 17 based on a signal output from the voltage sensor 17 during the measurement period Pm.

[0083] The acquisition unit 11 acquires the current measurement value CMtm by each current sensor 16 and the voltage measurement value VMtm by the voltage sensor 17 at each measurement time tm according to a predetermined measurement cycle Cm, generates measurement information including the acquired current measurement value CMtm and voltage measurement value VMtm, the measurement time tm, and the IDs of each of the four current sensors 16 and voltage sensors 17, and outputs the generated measurement information to the communication unit 14. The measurement time tm is an example of a target time. The measurement cycle Cm is, for example, 10 minutes.

[0084] The communication unit 14 is capable of performing power line communication with the collecting device 151 via the aggregation lines 2 and 5. Every time the communication unit 14 receives measurement information from the acquiring unit 11, the communication unit 14 includes the ID of the monitoring device 111 in the received measurement information and transmits the information to the collecting device 151 via the PCS 8.

[0085] Each time the collecting device 151 receives measurement information from each monitoring device 111 installed in the power collection unit 60 in the corresponding PCS unit 80, it generates collected information including the measurement information received from each monitoring device 111 and the ID of the collecting device 151, and transmits the generated collected information to the detection device 101.

[0086] The detection device 101 receives collected information from each collection device 151. For example, the detection device 101 performs a detection process to detect an abnormality in the power generation unit 78 in the solar power generation system 401 based on the received collected information.

[0087] [assignment] 5 again, in the photovoltaic power generation system 401, when the output of a solar cell 91 in a cluster 92 is reduced due to high resistance or the effect of shading, a bypass diode 93 in the cluster 92 is short-circuited, forming a detour for the generated current. The bypass diode 93 may lead to an open circuit failure if the short-circuited state continues. In this case, the solar cell 91 with high resistance or disconnection may generate heat and cause a fire. In order to prevent a fire in the cluster 92, a technology is desired that can detect a state in which the bypass diode 93 is maintained short-circuited for a long period of time as an abnormality in the power generation unit 78 in the photovoltaic power generation system 401 before the bypass diode 93 leads to an open circuit failure.

[0088] Except for lightning strikes, it is rare for multiple clusters 92 in one power generating unit 78 to fail simultaneously, and in many cases, a single cluster 92 fails. For example, in a power generating unit 78 in which 14 solar cell panels 79 are connected in series, if the bypass diode 93 in one cluster 92 is short-circuited, the power generation amount of the power generating unit 78 decreases by approximately 2.38%. Also, in a power generating unit 78 in which 20 solar cell panels 79 are connected in series, if the bypass diode 93 in one cluster 92 is short-circuited, the power generation amount of the power generating unit 78 decreases by approximately 1.67%.

[0089] In this way, since the rate of decrease in the amount of power generated by the power generation unit 78 when one bypass diode 93 is short-circuited is very small, it is not easy to detect a short circuit in the bypass diode 93 using the techniques described in Patent Documents 1 and 2. In addition, the short circuit in the bypass diode 93 when the output of the solar cell 91 decreases due to the influence of a shadow is temporary, and it is not easy to detect a state in which the short circuit in the bypass diode 93 has been maintained for a long period of time, distinguishing it from a temporary short circuit caused by the influence of a shadow.

[0090] Furthermore, with the technology described in Patent Document 2, it is not possible to detect an abnormality in the power generation section 78 unless output suppression is being performed.

[0091] Therefore, the detection device 101 according to the embodiment of the present disclosure solves the above problem by having the following configuration.

[0092] [Detection device configuration] FIG. 8 is a diagram illustrating a configuration of a detection device according to an embodiment of the present disclosure. With reference to FIG. 8, the detection device 101 includes an acquisition unit 31, a processing unit 32, and a storage unit 33. The acquisition unit 31 is an example of a performance value acquisition unit. The processing unit 32 is an example of a threshold acquisition unit, an example of a predicted value calculation unit, an example of a feature calculation unit, and an example of a determination unit. A part or all of the acquisition unit 31 and the processing unit 32 are realized, for example, by a processing circuit including one or more processors. The storage unit 33 is, for example, a non-volatile memory included in the processing circuit.

[0093] For example, some or all of the functions of the detection device 101 may be provided by cloud computing. That is, the detection device 101 may be configured by a plurality of cloud servers, etc. The detection device 101 may be built into the collection device 151, for example.

[0094] The storage unit 33 stores a correspondence table Tb1 indicating the correspondence between the ID of the collecting device 151, the ID of the monitoring device 111, the ID of the current sensor 16, the ID of the voltage sensor 17, the ID of the power generation unit 78, the ID of the solar cell unit 75, and the PCS 8. Hereinafter, the ID of the collecting device 151 will also be referred to as the "collecting device ID", the ID of the monitoring device 111 will also be referred to as the "monitoring device ID", the ID of the current sensor 16 will also be referred to as the "current sensor ID", the ID of the voltage sensor 17 will also be referred to as the "voltage sensor ID", the ID of the power generation unit 78 will also be referred to as the "power generation unit ID", and the ID of the solar cell unit 75 will also be referred to as the "solar cell unit ID".

[0095] The storage unit 33 also stores flag information FR indicating the correspondence between the power generation unit ID and flags F1, F2, and F3. The flag F1 is an output reduction flag indicating whether or not an output reduction corresponding to the power generation amount of the cluster 92 has occurred. The flag F2 is a tentative determination flag indicating whether or not there is a suspicion of an abnormality in the power generation unit 78. The flag F3 is an abnormality determination flag indicating whether or not an abnormality has occurred in the power generation unit 78. The flags F1, F2, and F3 in the flag information FR are initially off, and are switched between off and on by the processing unit 32.

[0096] The storage unit 33 also stores a lower limit threshold ThL and an upper limit threshold ThH as a threshold Th used in a determination process described later. The threshold Th is set in advance based on the magnitude of the output of the cluster 92 in the solar panel 79.

[0097] For example, the threshold value Th is set based on the output characteristics of a normal string, which is a power generation unit 78 including a solar panel 79 in which the output of the cluster 92 is normal, and the output characteristics of a reduced-output string, which is a power generation unit 78 including a solar panel 79 in which the output of the cluster 92 has decreased due to a short circuit of the bypass diode 93. Here, the reduced-output string is assumed to include one solar panel 79 in one of the three clusters 92 in which the bypass diode 93 is short-circuited.

[0098] Fig. 9 is a diagram showing an example of a method for setting the lower threshold value ThL stored in a memory unit in a detection device according to an embodiment of the present disclosure. In Fig. 9, the horizontal axis is voltage [V] and the vertical axis is power [W]. The solid lines in Fig. 9 indicate the PV curve Cpv1 of a normal string and the PV curve Cpv2 of a string with reduced output. The PV curves Cpv1 and Cpv2 are examples of power-voltage characteristics.

[0099] 9, the lower threshold ThL is set based on the output power Wa1 of the normal string at the voltage value Vp3 and the output power Wa2 of the reduced output string at the voltage value Vp3. More specifically, the lower threshold ThL is set to a value obtained by dividing the output power Wa2 corresponding to the voltage value Vp3 in the PV curve Cpv2 by the output power Wa1 corresponding to the voltage value Vp3 in the PV curve Cpv1. Here, the voltage value Vp3 is a voltage value that is greater than the voltage value Vp1 corresponding to the peak power in the PV curve Cpv1 and smaller than a voltage value Vp4 obtained by adding the absolute value of the difference Vdif between the voltage value Vp2 corresponding to the peak power in the PV curve Cpv2 and the voltage value Vp1 to the voltage value Vp1. The voltage value Vp1 is an example of a first voltage value, the voltage value Vp2 is an example of a second voltage value, and the voltage value Vp3 is an example of a third voltage value.

[0100] 10 is a diagram showing an example of a method for setting the upper threshold value ThH stored in the storage unit of the detection device according to the embodiment of the present disclosure, which can be viewed in the same way as in FIG.

[0101] 10, the upper threshold ThH is set based on the output power Wb1 of the normal string at the voltage value Vp2 and the output power Wb2 of the reduced output string at the voltage value Vp2. More specifically, the upper threshold ThH is set to a value obtained by dividing the output power Wb2 corresponding to the voltage value Vp2 on the PV curve Cpv2 by the output power Wb1 corresponding to the voltage value Vp2 on the PV curve Cpv1.

[0102] The threshold value Th may be set in advance based on the magnitude of the power generation current of the cluster 92. In other words, the lower limit threshold value ThL and the upper limit threshold value ThH may be set based on the IV curve of a normal string and the IV curve of a reduced output string, instead of the PV curves Cpv1 and Cpv2.

[0103] (Acquisition Department) 8 again, the acquisition unit 31 acquires an actual value of the output of the power generation unit 78. For example, the acquisition unit 31 acquires a current measurement value CMtm as the actual value of the output of the power generation unit 78. More specifically, the acquisition unit 31 receives collected information from each collection device 151 via VPN. The acquisition unit 31 stores the received collected information in the storage unit 33.

[0104] For example, when the end time of the measurement period Pm of the day arrives, the acquisition unit 31 generates current data Dc, which is time-series data indicating the correspondence between the measurement time tm in the measurement period Pm of the day and the current measurement value CMtm at the measurement time tm, based on the correspondence table Tb1 in the storage unit 33 and the collected information received during the measurement period Pm of the day. The acquisition unit 31 generates the current data Dc for each power generation unit 78, and stores the generated current data Dc in the storage unit 33 in association with the power generation unit ID and the solar cell unit ID.

[0105] Also, for example, the acquisition unit 31 acquires an actual power value PMtm, which is an actual value of the power generation power P of the power generation unit 78, as an actual value of the output of the power generation unit 78. More specifically, the acquisition unit 31 calculates the actual power value PMtm of the power generation unit 78 at the measurement time tm by multiplying the current measurement value CMtm at the measurement time tm by the voltage measurement value VMtm at the measurement time tm, based on the correspondence table Tb1 in the storage unit 33 and the collected information received during the measurement period Pm of the day. The acquisition unit 31 calculates the actual power value PMtm for each measurement time tm, and generates power data Dp, which is time-series data indicating the correspondence between the measurement time tm and the actual power value PMtm during the measurement period Pm of the day. The acquisition unit 31 generates the power data Dp for each power generation unit 78, and stores the generated power data Dp in the storage unit 33 in association with the power generation unit ID and the solar cell unit ID.

[0106] (Processing section) The processing unit 32 performs a detection process to detect an abnormality in the power generation unit 78. More specifically, the processing unit 32 detects a state in which the bypass diode 93 in the power generation unit 78 remains short-circuited for a long period of time as an abnormality in the power generation unit 78. For example, the processing unit 32 performs the detection process on all the power generation units 78 in the solar power generation system 401 in turn each day. The detection process on one power generation unit 78 by the processing unit 32 will be described in detail below. The power generation unit 78 that is the target of the detection process by the processing unit 32 is also referred to as the target power generation unit 78. The target power generation unit 78 is an example of a target string.

[0107] For example, in the detection process for the target power generation unit 78, the processing unit 32 executes pre-processing and main processing, which will be described later, or executes post-processing, which will be described later. More specifically, when the acquisition unit 31 stores the power data Dp of each power generation unit 78 for the day in the storage unit 33, the processing unit 32 checks the flags F1, F2, and F3 of the target power generation unit 78 by referring to the flag information FR in the storage unit 33. When the flags F1, F2, and F3 of the target power generation unit 78 are all off, the processing unit 32 does not execute post-processing for the target power generation unit 78, but executes pre-processing and main processing for the target power generation unit 78. On the other hand, when any one of the flags F1, F2, and F3 of the target power generation unit 78 is on, the processing unit 32 does not execute pre-processing and main processing for the target power generation unit 78, but executes post-processing for the target power generation unit 78.

[0108] (1) Pretreatment When the flags F1, F2, and F3 of the target power generation unit 78 are all off, the processing unit 32 acquires the power data Dp of the target power generation unit 78 during a calculation target period PT1 of a predetermined length and the power data Dp of the target power generation unit 78 on the current day from the storage unit 33. The calculation target period PT1 is a period that precedes a determination target period PT2, which will be described later, and is, for example, a period from two months before the current day to one month before the current day.

[0109] Furthermore, the processing unit 32 refers to the correspondence table Tb1 in the storage unit 33 to identify a reference power generation unit 78 that is a power generation unit 78 included in the same solar cell unit 75 as the target power generation unit 78. The solar cell unit 75 includes the target power generation unit 78 and reference power generation units 78A, 78B, and 78C that are reference power generation units 78. After identifying the reference power generation units 78A, 78B, and 78C, the processing unit 32 acquires, from the storage unit 33, power data Dp of the reference power generation units 78A, 78B, and 78C for the calculation target period PT1 and power data Dp of the reference power generation units 78A, 78B, and 78C for the current day. The reference power generation unit 78 is an example of a reference string.

[0110] (2) Main processing (2-1) Calculation of power prediction value PAtm The processing unit 32 calculates a predicted power value PAtm that is a predicted value of the generated power P of the target power generation unit 78 at the measurement time tm, based on the actual power value PMtm of the target power generation unit 78 and the actual power value PMtm of the reference power generation unit 78 .

[0111] For example, the processing unit 32 calculates the predicted power value PAtm by excluding the actual power value PMtm for a time period in which at least one of the actual power value PMtm of the target power generation unit 78, the actual power value PMtm of the reference power generation unit 78, and the ratio between the actual power value PMtm of the target power generation unit 78 and the actual power value PMtm of the reference power generation unit 78 satisfies a predetermined condition.

[0112] More specifically, the processing unit 32 detects, based on the power data Dp of the target power generation unit 78 and the reference power generation unit 78, low insolation time periods in which the actual power value PMtm corresponding to at least one of the target power generation unit 78 and the reference power generation unit 78 is 25% or less of the rated power, and high insolation time periods in which the actual power value PMtm corresponding to at least one of the target power generation unit 78 and the reference power generation unit 78 is 75% or more of the rated power. The processing unit 32 also detects, based on the power data Dp of the target power generation unit 78 and the reference power generation unit 78, shadow time periods in which the ratio of the actual power value PMtm corresponding to the target power generation unit 78 to the actual power value PMtm corresponding to at least one of the reference power generation units 78 is less than 80% or more than 120%. The processing unit 32 then sets the detected low insolation time periods, high insolation time periods, and shadow time periods as excluded time periods.

[0113] The processing unit 32 performs an extraction process to extract actual power values ​​PMtm for a time period excluding the excluded time period from the actual power values ​​PMtm included in the power data Dp of the target power generation unit 78 and the reference power generation unit 78.

[0114] The processing unit 32 uses the actual power values ​​PMtm for the calculation period PT1 out of the actual power values ​​PMtm extracted in the extraction process, according to a statistical method, a machine learning method, or a deep learning method, to generate an estimation formula FM which is a linear regression equation passing through the origin for calculating a predicted power value PAtm of the target power generation unit 78 at the measurement time tm on the day. The processing unit 32 calculates the predicted power value PAtm of the target power generation unit 78 on the day using the generated estimation formula FM and the actual power value PMtm of the reference power generation unit 78 on the day.

[0115] More specifically, the processing unit 32 generates an estimation formula FM1, which is an estimation formula FM, using the actual power values ​​PMtm of the target power generation unit 78 and the reference power generation unit 78A during the calculation target period PT1. Then, the processing unit 32 calculates a predicted power value PAtm1, which is the predicted power value PAtm of the target power generation unit 78 at the measurement time tm of the current day, by substituting the actual power value PMtm of the reference power generation unit 78A at the measurement time tm of the current day into the generated estimation formula FM1.

[0116] The processing unit 32 also generates an estimation formula FM2, which is an estimation formula FM, using the actual power values ​​PMtm of the target power generation unit 78 and the reference power generation unit 78B during the calculation target period PT1. The processing unit 32 then calculates a predicted power value PAtm2, which is the predicted power value PAtm of the target power generation unit 78 at the measurement time tm of the current day, by substituting the actual power value PMtm of the reference power generation unit 78B at the measurement time tm of the current day into the generated estimation formula FM2.

[0117] The processing unit 32 also generates an estimation formula FM3, which is an estimation formula FM, using the actual power values ​​PMtm of the target power generation unit 78 and the reference power generation unit 78C during the calculation target period PT1. The processing unit 32 then calculates a predicted power value PAtm3, which is the predicted power value PAtm of the target power generation unit 78 at the measurement time tm of the day, by substituting the actual power value PMtm of the reference power generation unit 78C at the measurement time tm of the day into the generated estimation formula FM3.

[0118] In addition, if all measurement times tm on that day are within the excluded time zone and the actual power value PMtm at the measurement time tm on that day is not extracted in the extraction process, the processing unit 32 does not generate the estimation formula FM or calculate the predicted power value PAtm of the target power generation unit 78 on that day.

[0119] (2-2) Calculation of the average feature value FVav The processing unit 32 calculates a feature value FV which is the ratio between the predicted power value PAtm of the target power generation unit 78 at the measurement time tm and the actual power value PMtm of the target power generation unit 78 at the measurement time tm.

[0120] More specifically, the processing unit 32 calculates the feature value FV1 for each measurement time tm by dividing the actual power value PMtm of the target power generation unit 78 at the measurement time tm on the day by the predicted power value PAtm1 of the target power generation unit 78 at the measurement time tm. The feature value FV1 is the feature value FV calculated using the actual power value PMtm of the reference power generation unit 78A.

[0121] Furthermore, the processing unit 32 calculates a feature value FV2 for each measurement time tm by dividing the actual power value PMtm of the target power generation unit 78 at the measurement time tm on the day by the predicted power value PAtm2 of the target power generation unit 78 at the measurement time tm. The feature value FV2 is a feature value FV calculated using the actual power value PMtm of the reference power generation unit 78B.

[0122] Furthermore, the processing unit 32 calculates a feature value FV3 for each measurement time tm by dividing the actual power value PMtm of the target power generation unit 78 at the measurement time tm on the day by the predicted power value PAtm3 of the target power generation unit 78 at the measurement time tm. The feature value FV3 is a feature value FV calculated using the actual power value PMtm of the reference power generation unit 78C.

[0123] After calculating the feature amount FV1 for each measurement time tm on the day, the processing unit 32 calculates an average feature amount FVav1, which is the average value of the feature amounts FV1 for each measurement time tm on the day, and stores the calculated average feature amount FVav1 in the storage unit 33.

[0124] In addition, after calculating the feature amount FV2 for each measurement time tm on the day, the processing unit 32 calculates an average feature amount FVav2, which is the average value of the feature amounts FV2 for each measurement time tm on the day, and stores the calculated average feature amount FVav2 in the memory unit 33.

[0125] Furthermore, after calculating the feature amount FV3 for each measurement time tm on the day, the processing unit 32 calculates an average feature amount FVav3 which is the average value of the feature amounts FV3 for each measurement time tm on the day, and stores the calculated average feature amount FVav3 in the storage unit 33. Hereinafter, each of the average feature amounts FVav1, FVav2, FVav3 will also be referred to as an average feature amount FVav.

[0126] If all measurement times tm on that day are within the excluded time zone and the actual power values ​​PMtm at the measurement times tm on that day are not extracted in the extraction process, the processing unit 32 does not calculate the average feature value FVav or store it in the memory unit 33.

[0127] (2-3) Judgment process The processing unit 32 calculates an evaluation value RV based on the feature amount FV at a plurality of measurement times tm in a predetermined length of a judgment period PT2, and performs judgment processing using the calculated evaluation value RV. The judgment period PT2 is, for example, the most recent seven days.

[0128] More specifically, when three or more average feature values ​​FVav1 calculated during the determination period PT2 are stored in the storage unit 33, the processing unit 32 calculates the average value of the average feature values ​​FVav1 calculated during the determination period PT2 as the evaluation value RV1. The evaluation value RV1 is an evaluation value RV calculated using the actual power value PMtm of the reference power generation unit 78A. Note that when fewer than three average feature values ​​FVav1 are calculated during the determination period PT2, the processing unit 32 does not calculate the evaluation value RV1.

[0129] Furthermore, when three or more average feature values ​​FVav2 calculated during the determination period PT2 are stored in the storage unit 33, the processing unit 32 calculates the average of the average feature values ​​FVav2 calculated during the determination period PT2 as the evaluation value RV2. The evaluation value RV2 is an evaluation value RV calculated using the actual power value PMtm of the reference power generation unit 78B. Note that when fewer than three average feature values ​​FVav2 are calculated during the determination period PT2, the processing unit 32 does not calculate the evaluation value RV2.

[0130] Furthermore, when three or more average feature values ​​FVav3 calculated during the determination period PT2 are stored in the storage unit 33, the processing unit 32 calculates the average of the average feature values ​​FVav3 calculated during the determination period PT2 as the evaluation value RV3. The evaluation value RV3 is an evaluation value RV calculated using the actual power value PMtm of the reference power generation unit 78C. Note that when fewer than three average feature values ​​FVav3 are calculated during the determination period PT2, the processing unit 32 does not calculate the evaluation value RV3.

[0131] After calculating the evaluation value RV, the processing unit 32 acquires a lower limit threshold ThL and an upper limit threshold ThH from the storage unit 33. Then, the processing unit 32 performs a determination process for determining an output reduction corresponding to the power generation amount of the cluster 92 in the target power generation unit 78, based on a comparison result between the calculated evaluation value RV and the acquired lower limit threshold ThL and upper limit threshold ThH.

[0132] For example, if more than half of the multiple evaluation values ​​RV calculated for each reference power generation unit 78 are greater than or equal to the lower threshold value ThL and less than or equal to the upper threshold value ThH, the processing unit 32 determines that an output reduction equivalent to the power generation amount of the cluster 92 has occurred in the target power generation unit 78.

[0133] More specifically, when any two or more of the evaluation values ​​RV1, RV2, RV3 are equal to or greater than the lower threshold ThL and equal to or less than the upper threshold ThH, the processing unit 32 determines that an output reduction equivalent to the power generation amount of the cluster 92 has occurred in the target power generation unit 78. On the other hand, when there are less than two evaluation values ​​RV that are equal to or greater than the lower threshold ThL and equal to or less than the upper threshold ThH, the processing unit 32 determines that an output reduction equivalent to the power generation amount of the cluster 92 has not occurred in the target power generation unit 78.

[0134] When the processing unit 32 determines that an output reduction equivalent to the power generation amount of the cluster 92 has occurred in the target power generation unit 78, it switches the flag F1 of the target power generation unit 78 in the flag information FR in the storage unit 33 from OFF to ON.

[0135] (3) Post-processing (3-1) Post-processing of flag F1 When the flag F1 of the target power generation unit 78 is on, the processing unit 32 acquires the power data Dp of the target power generation unit 78 for a predetermined length of monitoring period PT3 from the storage unit 33. The monitoring period PT3 is, for example, a period from two months before the current day to the current day.

[0136] In addition, the processing unit 32 refers to the flag information FR and the correspondence table Tb1 in the memory unit 33 and obtains from the memory unit 33 reference power data DpR, which is the power data Dp during the monitoring period PT3, for each of the multiple power generation units 78 that are connected to the same PCS8 as the target power generation unit 78 and whose flags F1, F2, and F3 are all off.

[0137] The processing unit 32 uses the acquired plurality of reference power data DpR to calculate a representative value of the outputs of the plurality of power generation units 78 connected to the same PCS 8 as the target power generation unit 78. More specifically, the processing unit 32 calculates, as the representative value, a median MEDtm of the actual power values ​​PMtm included in the plurality of reference power data DpR for each measurement time tm in the monitoring period PT3. Note that the processing unit 32 may be configured to calculate, instead of the median MEDtm, other values ​​such as the maximum value, the first quartile, or the third quartile of the actual power values ​​PMtm included in the plurality of reference power data DpR for each measurement time tm in the monitoring period PT3 as the representative value.

[0138] The processing unit 32 determines whether or not to validate the judgment result that the output of the target power generation unit 78 has decreased in the judgment process, based on the output ratio R1tm, which is the ratio between the median value MEDtm and the actual power value PMtm of the target power generation unit 78.

[0139] More specifically, the processing unit 32 calculates an output ratio R1tm, which is a value obtained by dividing the actual power value PMtm included in the power data Dp of the target power generation unit 78 by the median value MEDtm, for each measurement time tm in the monitoring period PT3. The processing unit 32 performs a determination process to determine whether or not to keep the flag F1 of the target power generation unit 78 on, based on the output ratio R1tm for each measurement time tm in the monitoring period PT3.

[0140] (3-1-1) Judgment process 1 When the difference between a plurality of output ratios R1tm corresponding to a plurality of measurement times tm and an output ratio R1tm within a predetermined range among the plurality of output ratios R1tm is a predetermined value or more, the processing unit 32 invalidates the judgment result that the output of the target power generation unit 78 has decreased in the judgment process.

[0141] More specifically, the processing unit 32 classifies all the calculated output ratios R1tm into output ratios R1Aam, which are the output ratios R1tm in the numerical range RgA corresponding to the measurement time tm in the morning, output ratios R1Apm, which are the output ratios R1tm in the numerical range RgA corresponding to the measurement time tm in the afternoon, output ratios R1Bam, which are the output ratios R1tm in the numerical range RgB corresponding to the measurement time tm in the morning, and output ratios R1Bpm, which are the output ratios R1tm in the numerical range RgB corresponding to the measurement time tm in the afternoon. For example, the numerical range RgA is a range of 0.1 or more and 10 or less. Also, for example, the numerical range RgB is a range of 0.8 or more and 1.2 or less. The processing unit 32 classifies the output ratios R1tm corresponding to the measurement time tm in the morning in the overlapping range of the numerical range RgA and the numerical range RgB into both the output ratios R1Aam and the output ratios R1Bam. Furthermore, the output ratio R1tm corresponding to the measurement time tm in the afternoon in the overlapping range of the numerical range RgA and the numerical range RgB is classified into both the output ratio R1Apm and the output ratio R1Bpm.

[0142] The processing unit 32 performs an adding process and a pruning process on the classified multiple output ratios R1Aam, each time the output ratio R1Aam corresponds to the measurement time tm for one day, to extract 100 output ratios R1A1, which are output ratios R1Aam from which abnormal values ​​and outliers have been removed.

[0143] In addition, the processing unit 32 performs an adding process and a pruning process on the classified multiple output ratios R1Apm, each time the output ratio R1Apm corresponds to the measurement time tm for one day, to extract 100 output ratios R1A2, which are output ratios R1Apm from which abnormal values ​​and outliers have been removed.

[0144] In addition, the processing unit 32 performs an adding process and a pruning process on the classified multiple output ratios R1Bam, each time the output ratio R1Bam corresponds to the measurement time tm for one day, to extract 100 output ratios R1B1, which are output ratios R1Bam from which abnormal values ​​and outliers have been removed.

[0145] In addition, the processing unit 32 performs an adding process and a pruning process on the classified multiple output ratios R1Bpm, each of which corresponds to the measurement time tm for one day, to extract 100 output ratios R1B2, which are output ratios R1Bpm from which abnormal values ​​and outliers have been removed.

[0146] The processing unit 32 calculates an absolute value AB of the difference between the average value of the extracted 100 output ratios R1A1 and 100 output ratios R1A2 and the average value of the extracted 100 output ratios R1B1 and 100 output ratios R1B2, and compares the calculated absolute value AB with a predetermined threshold value Thx1. If the absolute value AB is equal to or greater than the threshold value Thx1, the processing unit 32 determines that the flag F1 of the target power generation unit 78 is turned on due to a drop in output caused by a shadow, and switches the flag F1 of the target power generation unit 78 from on to off. On the other hand, if the absolute value AB is less than the threshold value Thx1, the processing unit 32 maintains the flag F1 of the target power generation unit 78 on.

[0147] (3-1-2) Judgment process 2 If the rate of decline DR1 of the output ratios R1A1, R1A2 during a predetermined period prior to the measurement time tm of the actual power value PMtm used to calculate the evaluation value RV is equal to or greater than a predetermined value, the processing unit 32 invalidates the judgment result that the output of the target power generation unit 78 has declined in the judgment process.

[0148] More specifically, the processing unit 32 calculates the rate of decrease DR1 of the output ratios R1A1, R1A2 during the period from one month before the day on which the processing unit 32 determined that the target power generation unit 78 had experienced an output decrease and turned on the flag F1 of the target power generation unit 78, and compares the calculated rate of decrease DR1 with a predetermined threshold value Thx2. The threshold value Thx2 is, for example, 5%. If the rate of decrease DR1 is equal to or greater than the threshold value Thx2, the processing unit 32 determines that the flag F1 of the target power generation unit 78 has been turned on due to an output decrease caused by a shadow, and switches the flag F1 of the target power generation unit 78 from on to off. On the other hand, if the rate of decrease DR1 is less than the threshold value Thx2, the processing unit 32 maintains the flag F1 of the target power generation unit 78 on.

[0149] (3-1-3) Judgment process 3 If the rate of decline DR2 of the output ratios R1A1, R1A2 during a predetermined period of time after the measurement time tm of the actual power value PMtm used to calculate the evaluation value RV is equal to or greater than a predetermined value, the processing unit 32 invalidates the judgment result that the output of the target power generation unit 78 has declined in the judgment process.

[0150] More specifically, the processing unit 32 calculates a rate of decrease DR2 of the output ratios R1A1, R1A2 during the period from the day when the target power generation unit 78's output decrease was determined and the flag F1 of the target power generation unit 78 was turned on to the current day, and compares the calculated rate of decrease DR2 with a predetermined threshold value Thx3. The threshold value Thx3 is, for example, 2%. If the rate of decrease DR2 is equal to or greater than the threshold value Thx3, the processing unit 32 determines that the target power generation unit 78's flag F1 has been turned on due to an output decrease caused by a shadow, and switches the flag F1 of the target power generation unit 78 from on to off. On the other hand, if the rate of decrease DR2 is less than the threshold value Thx3, the processing unit 32 maintains the flag F1 of the target power generation unit 78 on.

[0151] (3-1-4) Judgment process 4 If the increase rate IR1 of the output ratios R1B1, R1B2 during a predetermined period of time after the measurement time tm of the actual power value PMtm used to calculate the evaluation value RV is greater than or equal to a predetermined value, the processing unit 32 invalidates the judgment result that the output of the target power generation unit 78 has decreased in the judgment process.

[0152] More specifically, the processing unit 32 calculates an increase rate IR1 of the output ratios R1B1, R1B2 during the period from the day when the target power generation unit 78 was determined to have experienced an output decrease and the flag F1 of the target power generation unit 78 was turned on to the current day, and compares the calculated increase rate IR1 with a predetermined threshold value Thx4. The threshold value Thx4 is, for example, 2%. If the increase rate IR1 is equal to or greater than the threshold value Thx4, the processing unit 32 determines that the output decrease corresponding to the amount of power generated by the cluster 92 has been resolved in the target power generation unit 78, and switches the flag F1 of the target power generation unit 78 from on to off. On the other hand, if the increase rate IR1 is less than the threshold value Thx4, the processing unit 32 maintains the flag F1 of the target power generation unit 78 on.

[0153] (3-1-5) Switching from flag F1 to flag F2 If the processing unit 32 determines in the determination process that the flag F1 of the target power generation unit 78 remains on, it counts the period of time that has elapsed since the day that the flag F1 of the target power generation unit 78 was turned on.

[0154] When the period of time that has elapsed since the day flag F1 of the target power generation unit 78 was turned on reaches one month, the processing unit 32 determines that there is a possibility that an abnormality has occurred in the target power generation unit 78, switches off flag F1 of the target power generation unit 78, and switches on flag F2 of the target power generation unit 78. Then, the processing unit 32 performs a notification process to notify the administrator of the photovoltaic power generation system 401 of the determination result that there is a possibility that an abnormality has occurred in the target power generation unit 78.

[0155] On the other hand, if the period of time that has elapsed since the flag F1 of the target power generation unit 78 was turned on is less than one month, the processing unit 32 keeps the flag F1 of the target power generation unit 78 on and ends the post-processing.

[0156] (3-2) Post-processing for flag F2 When the flag F2 of the target power generation unit 78 is on, the processing unit 32 acquires from the storage unit 33 the power data Dp of the target power generation unit 78 during the monitoring period PT3.

[0157] In addition, the processing unit 32 refers to the flag information FR and the correspondence table Tb1 in the memory unit 33 and obtains from the memory unit 33 reference power data DpR, which is the power data Dp during the monitoring period PT3, for each of the multiple power generation units 78 that are connected to the same PCS8 as the target power generation unit 78 and whose flags F1, F2, and F3 are all off.

[0158] The processing unit 32 uses the acquired plurality of reference power data DpR to calculate a representative value of the outputs of the plurality of power generation units 78 connected to the same PCS 8 as the target power generation unit 78. More specifically, the processing unit 32 calculates, as the representative value, a median value MEDtm of the actual power values ​​PMtm included in the plurality of reference power data DpR for each measurement time tm during the monitoring period PT3.

[0159] The processing unit 32 detects an abnormality in the target power generating unit 78 based on the output ratio R2tm, which is the ratio between the median value MEDtm and the actual power value PMtm of the target power generating unit 78.

[0160] More specifically, the processing unit 32 calculates an output ratio R2tm, which is a value obtained by dividing the actual power value PMtm included in the power data Dp of the target power generation unit 78 by the median value MEDtm, for each measurement time tm in the monitoring period PT3. The processing unit 32 performs a determination process to determine whether or not to keep the flag F2 of the target power generation unit 78 on, based on the output ratio R2tm for each measurement time tm in the monitoring period PT3.

[0161] (3-2-1) Judgment process 5 The processing unit 32 classifies all the calculated output ratios R2tm into output ratios R2Aam, which are the output ratios R2tm in the numerical range RgA corresponding to the morning measurement time tm, output ratios R2Apm, which are the output ratios R2tm in the numerical range RgA corresponding to the afternoon measurement time tm, output ratios R2Bam, which are the output ratios R2tm in the numerical range RgB corresponding to the morning measurement time tm, and output ratios R2Bpm, which are the output ratios R2tm in the numerical range RgB corresponding to the afternoon measurement time tm.

[0162] The processing unit 32 performs an addition process and a pruning process on the classified multiple output ratios R2Bam, each time the output ratio R2Bam corresponding to the measurement time tm for one day, to extract 100 output ratios R2B1 that are output ratios R2Bam from which abnormal values ​​and outliers have been removed. The processing unit 32 also performs an addition process and a pruning process on the classified multiple output ratios R2Bpm, each time the output ratio R2Bpm corresponding to the measurement time tm for one day, to extract 100 output ratios R2B2 that are output ratios R2Bpm from which abnormal values ​​and outliers have been removed.

[0163] If the increase rate IR2 of the output ratios R2B1, R2B2 during a specified period of time after the day on which the flag F2 of the target power generation unit 78 was turned on is greater than or equal to a specified value, the processing unit 32 invalidates the judgment result that there is a possibility that an abnormality has occurred in the target power generation unit 78.

[0164] More specifically, the processing unit 32 calculates the increase rate IR2 of the output ratios R2B1, R2B2 during the period from the day the flag F2 of the target power generation unit 78 was turned on to the current day, and compares the calculated increase rate IR2 with a predetermined threshold value Thx5. The threshold value Thx5 is, for example, 2%. If the increase rate IR2 is equal to or greater than the threshold value Thx5, the processing unit 32 determines that the output reduction equivalent to the power generation amount of the cluster 92 has been resolved in the target power generation unit 78, and switches the flag F2 of the target power generation unit 78 from on to off. On the other hand, if the increase rate IR2 is less than the threshold value Thx5, the processing unit 32 maintains the flag F2 of the target power generation unit 78 on.

[0165] (3-2-2) Switching from flag F2 to flag F3 If the processing unit 32 determines in the determination process that the flag F2 of the target power generation unit 78 remains on, it counts the period of time that has elapsed since the day the flag F2 of the target power generation unit 78 was turned on.

[0166] When the period of time that has elapsed since the day when the flag F2 of the target power generation unit 78 was turned on reaches one month, the processing unit 32 determines that an abnormality has occurred in the target power generation unit 78 because the short circuit of the bypass diode 93 has been maintained for a long period of time, switches off the flag F2 of the target power generation unit 78, and switches on the flag F3 of the target power generation unit 78. Then, the processing unit 32 performs a notification process to notify the administrator of the photovoltaic power generation system 401 of the determination result that an abnormality has occurred in the target power generation unit 78.

[0167] On the other hand, if the period of time that has elapsed since the flag F2 of the target power generation unit 78 was turned on is less than one month, the processing unit 32 keeps the flag F2 of the target power generation unit 78 on and ends the post-processing.

[0168] (3-3) Post-processing for flag F3 When the flag F3 of the target power generation unit 78 is on, the processing unit 32 acquires from the storage unit 33 the power data Dp of the target power generation unit 78 during the monitoring period PT3.

[0169] In addition, the processing unit 32 refers to the flag information FR and the correspondence table Tb1 in the memory unit 33 and obtains from the memory unit 33 reference power data DpR, which is the power data Dp during the monitoring period PT3, for each of the multiple power generation units 78 that are connected to the same PCS8 as the target power generation unit 78 and whose flags F1, F2, and F3 are all off.

[0170] The processing unit 32 uses the acquired plurality of reference power data DpR to calculate a representative value of the outputs of the plurality of power generation units 78 connected to the same PCS 8 as the target power generation unit 78. More specifically, the processing unit 32 calculates, as the representative value, a median value MEDtm of the actual power values ​​PMtm included in the plurality of reference power data DpR for each measurement time tm during the monitoring period PT3.

[0171] The processing unit 32 determines whether the abnormality of the target power generating unit 78 is maintained based on the output ratio R3tm, which is the ratio between the median value MEDtm and the actual power value PMtm of the target power generating unit 78.

[0172] More specifically, the processing unit 32 calculates an output ratio R3tm, which is a value obtained by dividing the actual power value PMtm included in the power data Dp of the target power generation unit 78 by the median value MEDtm, for each measurement time tm in the monitoring period PT3. The processing unit 32 performs a determination process to determine whether or not to keep the flag F3 of the target power generation unit 78 on, based on the output ratio R3tm for each measurement time tm in the monitoring period PT3.

[0173] (3-3-1) Judgment process 6 The processing unit 32 classifies all the calculated output ratios R3tm into output ratios R3Aam, which are the output ratios R3tm in the numerical range RgA corresponding to the morning measurement time tm, output ratios R3Apm, which are the output ratios R3tm in the numerical range RgA corresponding to the afternoon measurement time tm, output ratios R3Bam, which are the output ratios R3tm in the numerical range RgB corresponding to the morning measurement time tm, and output ratios R3Bpm, which are the output ratios R3tm in the numerical range RgB corresponding to the afternoon measurement time tm.

[0174] The processing unit 32 performs addition and pruning processes on the classified multiple output ratios R3Bam, one by one, for the output ratios R3Bam corresponding to the measurement time tm for one day, to extract 100 output ratios R3B1 that are output ratios R3Bam from which abnormal values ​​and outliers have been removed. The processing unit 32 also performs addition and pruning processes on the classified multiple output ratios R3Bpm, one by one, for the output ratios R3Bpm corresponding to the measurement time tm for one day, to extract 100 output ratios R3B2 that are output ratios R3Bpm from which abnormal values ​​and outliers have been removed.

[0175] If the increase rate IR3 of the output ratios R3B1, R3B2 during a predetermined period after the day the flag F3 of the target power generating unit 78 was turned on is equal to or greater than a predetermined value, the processing unit 32 determines that the abnormality of the target power generating unit 78 has been resolved.

[0176] More specifically, the processing unit 32 calculates the increase rate IR3 of the output ratios R3B1, R3B2 during the period from the day the flag F3 of the target power generation unit 78 was turned on to the current day, and compares the calculated increase rate IR3 with a predetermined threshold value Thx6. The threshold value Thx6 is, for example, 2%. If the increase rate IR3 is equal to or greater than the threshold value Thx6, the processing unit 32 determines that the abnormality of the target power generation unit 78 has been resolved, and switches the flag F3 of the target power generation unit 78 from on to off. On the other hand, if the increase rate IR3 is less than the threshold value Thx6, the processing unit 32 maintains the flag F3 of the target power generation unit 78 on.

[0177] [Operation flow] FIG. 11 is a flowchart defining an example of an operation procedure when the detection device according to the embodiment of the present disclosure performs the detection process.

[0178] Referring to FIG. 11, first, the detection device 101 waits for the arrival of collected information and the end time of the measurement period Pm (NO in step S11 and NO in step S12), and upon receiving collected information from each collection device 151 (YES in step S11), it stores the received collected information in the memory unit 33 (step S12).

[0179] Next, the detection device 101 waits for the arrival of new collected information and the end time of the measurement period Pm (NO in step S11 and NO in step S12), and when the end time of the measurement period Pm for that day arrives (YES in step S12), it generates current data Dc and power data Dp for each power generation unit 78 based on the collected information for that day received from each collection device 151 (step S14).

[0180] Next, the detection device 101 refers to the flag information FR in the storage unit 33 to check the flags F1, F2, and F3 of the target power generation unit 78 (step S15).

[0181] Next, if the flags F1, F2, and F3 of the target power generation unit 78 are all off (YES in step S16), the detection device 101 performs pre-processing and main processing (step S17).

[0182] On the other hand, if any one of the flags F1, F2, F3 of the target power generation unit 78 is on (NO in step S16), the detection device 101 performs post-processing (step S18).

[0183] Next, the detection device 101 repeats the processes from steps S14 to S18 until the detection process is completed for all power generation units 78 in the solar power generation system 401 (NO in step S19), and when the detection process is completed for all power generation units 78 in the solar power generation system 401 (YES in step S19), it waits for the arrival of new collected information and the end time of the measurement period Pm (NO in step S11 and NO in step S12).

[0184] 12 is a flowchart showing an example of an operation procedure when the detection device according to the embodiment of the present disclosure performs a main process. FIG 12 shows details of the main process in step S17 in FIG 11.

[0185] Referring to FIG. 12, first, the detection device 101 sets an exclusion time period based on the power data Dp of the target power generation unit 78 and the reference power generation unit 78, and performs an extraction process to extract the actual power values ​​PMtm for a time period excluding the exclusion time period from the actual power values ​​PMtm included in the power data Dp of the target power generation unit 78 and the reference power generation unit 78 (step S21).

[0186] Next, the detection device 101 calculates the predicted power value PAtm of the target power generation unit 78 at the measurement time tm. More specifically, the detection device 101 generates estimation formulas FM1, FM2, and FM3 using the actual power value PMtm for the calculation target period PT1 among the actual power values ​​PMtm extracted in the extraction process. The detection device 101 then calculates the predicted power values ​​PAtm1, PAtm2, and PAtm3 by substituting the actual power values ​​PMtm of the reference power generation units 78A, 78B, and 78C at the measurement time tm of the day into the estimation formulas FM1, FM2, and FM3, respectively (step S22).

[0187] Next, the detection device 101 calculates a feature amount FV which is a ratio between the predicted power value PAtm of the target power generation unit 78 at the measurement time tm and the actual power value PMtm of the target power generation unit 78 at the measurement time tm. More specifically, the detection device 101 calculates the feature amounts FV1, FV2, and FV3 by dividing the actual power value PMtm of the target power generation unit 78 at the measurement time tm on the day by the predicted power values ​​PAtm1, PAtm2, and PAtm3 of the target power generation unit 78 at the measurement time tm (step S23).

[0188] Next, the detection device 101 calculates the average feature amount FVav. More specifically, the detection device 101 calculates an average feature amount FVav1 of the feature amount FV1 at each measurement time tm on that day, an average feature amount FVav2 of the feature amount FV2 at each measurement time tm on that day, and an average feature amount FVav3 of the feature amount FV3 at each measurement time tm on that day, and stores the calculated average feature amounts FVav1, FVav2, and FVav3 in the storage unit 33 (step S24).

[0189] Next, the detection device 101 calculates an evaluation value RV based on the feature amounts FV at multiple measurement times tm in the determination period PT2. More specifically, the detection device 101 calculates the average value of the average feature amounts FVav1 calculated in the determination period PT2 as the evaluation value RV1, calculates the average value of the average feature amounts FVav2 calculated in the determination period PT2 as the evaluation value RV2, and calculates the average value of the average feature amounts FVav3 calculated in the determination period PT2 as the evaluation value RV3 (step S25).

[0190] Next, the detection device 101 performs a determination process based on a comparison result between the calculated evaluation value RV and the threshold value Th. More specifically, the detection device 101 compares the evaluation value RV with a lower threshold value ThL and an upper threshold value ThH, which are the threshold values ​​Th (step S26).

[0191] Next, if less than half of the evaluation values ​​RV are greater than or equal to the lower threshold ThL and less than or equal to the upper threshold ThH (NO in step S27), the detection device 101 determines that no output reduction equivalent to the power generation amount of the cluster 92 has occurred in the target power generation unit 78 (step S28), and terminates the main processing.

[0192] On the other hand, if more than half of the multiple evaluation values ​​RV are greater than or equal to the lower threshold ThL and less than or equal to the upper threshold ThH (YES in step S27), the detection device 101 determines that an output reduction equivalent to the power generation amount of the cluster 92 has occurred in the target power generation unit 78 (step S29).

[0193] Next, the detection device 101 switches the flag F1 of the target power generation unit 78 in the flag information FR in the storage unit 33 from OFF to ON (step S30), and ends the main process.

[0194] Fig. 13 is a flowchart showing an example of an operation procedure when the detection device according to the embodiment of the present disclosure performs post-processing. Fig. 13 shows details of step S18 in Fig. 11 when flag F1 is on in step S16 in Fig. 11.

[0195] 13, first, the detection device 101 acquires the power data Dp of the target power generation unit 78 for the monitoring period PT3 from the storage unit 33 (step S41).

[0196] Next, the detection device 101 acquires from the memory unit 33 reference power data DpR, which is the power data Dp during the monitoring period PT3, for each of the multiple power generation units 78 that are connected to the same PCS 8 as the target power generation unit 78 and whose flags F1, F2, and F3 are all off (step S42).

[0197] Next, the detection device 101 calculates the median value MEDtm for each measurement time tm in the monitoring period PT3 using the multiple acquired reference power data DpR (step S43).

[0198] Next, the detection device 101 calculates an output ratio R1tm, which is a value obtained by dividing the actual power value PMtm included in the power data Dp of the target power generation unit 78 by the median value MEDtm, for each measurement time tm in the monitoring period PT3 (step S44).

[0199] Next, the detection device 101 classifies all the calculated output ratios R1tm into output ratios R1Aam, which are the output ratios R1tm in the numerical range RgA corresponding to the measurement time tm in the morning, output ratios R1Apm, which are the output ratios R1tm in the numerical range RgA corresponding to the measurement time tm in the afternoon, output ratios R1Bam, which are the output ratios R1tm in the numerical range RgB corresponding to the measurement time tm in the morning, and output ratios R1Bpm, which are the output ratios R1tm in the numerical range RgB corresponding to the measurement time tm in the afternoon.Then, the detection device 101 extracts 100 output ratios R1A1, 100 output ratios R1A2, 100 output ratios R1B1, and 100 output ratios R1B2 from which abnormal values ​​and outliers have been removed (step S45).

[0200] Next, if the absolute value AB of the difference between the average value of the output ratios R1A1, R1A2 and the average value of the output ratios R1B1, R1B2 is greater than or equal to the threshold value Thx1 (NO in step S46), if the rate of decline DR1 of the output ratios R1A1, R1A2 in the period from one month before the day on which the flag F1 of the target power generation unit 78 was turned on to that day is greater than or equal to the threshold value Thx2 (NO in step S47), if the rate of decline DR2 of the output ratios R1A1, R1A2 in the period from the day on which the flag F1 of the target power generation unit 78 was turned on to the current day is greater than or equal to the threshold value Thx3 (NO in step S48), or if the rate of increase IR1 of the output ratios R1B1, R1B2 in the period from the day on which the flag F1 of the target power generation unit 78 was turned on to the current day is greater than or equal to the threshold value Thx4 (NO in step S49), the detection device 101 switches the flag F1 of the target power generation unit 78 from on to off (step S50) and terminates the post-processing for the target power generation unit 78.

[0201] On the other hand, if the absolute value AB is less than the threshold value Thx1 (YES in step S46), the decrease rate DR1 is less than the threshold value Thx2 (YES in step S47), the decrease rate DR2 is less than the threshold value Thx3 (YES in step S48), and the increase rate IR1 is less than the threshold value Thx4 (YES in step S49), the detection device 101 keeps the flag F1 of the target power generation unit 78 on (step S51).

[0202] Next, if the period of time that has elapsed since the flag F1 of the target power generation unit 78 was turned on is less than one month (NO in step S52), the detection device 101 ends the post-processing for the target power generation unit 78.

[0203] On the other hand, if the period of time that has elapsed since the flag F1 of the target power generation unit 78 was turned on reaches one month (YES in step S52), the detection device 101 determines that there is a possibility that an abnormality has occurred in the target power generation unit 78 (step S53).

[0204] Next, the detection device 101 switches off the flag F1 of the target power generation unit 78, and switches on the flag F2 of the target power generation unit 78 (step S54).

[0205] Next, the detection device 101 performs notification processing to notify the administrator of the solar power generation system 401 of the determination result that there is a possibility that an abnormality has occurred in the target power generation unit 78 (step S55), and terminates the post-processing for the target power generation unit 78.

[0206] Fig. 14 is a flowchart showing an example of an operation procedure when the detection device according to the embodiment of the present disclosure performs post-processing. Fig. 14 shows details of step S18 in Fig. 11 when flag F2 is on in step S16 in Fig. 11.

[0207] 14, first, the detection device 101 acquires the power data Dp of the target power generation unit 78 for the monitoring period PT3 from the storage unit 33 (step S61).

[0208] Next, the detection device 101 acquires from the memory unit 33 reference power data DpR, which is the power data Dp during the monitoring period PT3, for each of the multiple power generation units 78 that are connected to the same PCS 8 as the target power generation unit 78 and whose flags F1, F2, and F3 are all off (step S62).

[0209] Next, the detection device 101 calculates the median value MEDtm for each measurement time tm in the monitoring period PT3 using the multiple acquired reference power data DpR (step S63).

[0210] Next, the detection device 101 calculates the output ratio R2tm, which is the value obtained by dividing the actual power value PMtm included in the power data Dp of the target power generation unit 78 by the median value MEDtm, for each measurement time tm in the monitoring period PT3 (step S64).

[0211] Next, the detection device 101 classifies all the calculated output ratios R2tm into output ratios R2Aam, which are the output ratios R2tm in the numerical range RgA corresponding to the measurement time tm in the morning, output ratios R2Apm, which are the output ratios R2tm in the numerical range RgA corresponding to the measurement time tm in the afternoon, output ratios R2Bam, which are the output ratios R2tm in the numerical range RgB corresponding to the measurement time tm in the morning, and output ratios R2Bpm, which are the output ratios R2tm in the numerical range RgB corresponding to the measurement time tm in the afternoon.Then, the detection device 101 extracts 100 output ratios R2B1 and 100 output ratios R2B2 from which abnormal values ​​and outliers have been removed (step S65).

[0212] Next, if the increase rate IR2 of the output ratios R2B1, R2B2 during the period from the day the flag F2 of the target power generation unit 78 was turned on to the current day is greater than or equal to the threshold value Thx5 (NO in step S66), the detection device 101 switches the flag F2 of the target power generation unit 78 from on to off (step S67) and terminates the post-processing for the power generation unit 78.

[0213] On the other hand, if the increase rate IR2 is less than the threshold value Thx5 (YES in step S66), the detection device 101 keeps the flag F2 of the target power generation unit 78 ON (step S68).

[0214] Next, if the period of time that has elapsed since the flag F2 of the target power generation unit 78 was turned on is less than one month (NO in step S69), the detection device 101 ends the post-processing for the target power generation unit 78.

[0215] On the other hand, if one month has elapsed since the flag F2 of the target power generation unit 78 was turned on (YES in step S69), the detection device 101 detects an abnormality in the target power generation unit 78. In other words, the detection device 101 determines that an abnormality has occurred in the target power generation unit 78 (step S70).

[0216] Next, the detection device 101 switches off the flag F2 of the target power generation unit 78, and switches on the flag F3 of the target power generation unit 78 (step S71).

[0217] Next, the detection device 101 performs a notification process to notify the administrator of the photovoltaic power generation system 401 of the determination result that an abnormality has occurred in the target power generation unit 78 (step S72), and ends the post-processing for the target power generation unit 78.

[0218] Fig. 15 is a flowchart showing an example of an operation procedure when the detection device according to the embodiment of the present disclosure performs post-processing. Fig. 15 shows details of step S18 in Fig. 11 when flag F3 is on in step S16 in Fig. 11.

[0219] 15, first, the detection device 101 acquires the power data Dp of the target power generation unit 78 for the monitoring period PT3 from the storage unit 33 (step S81).

[0220] Next, the detection device 101 acquires from the memory unit 33 reference power data DpR, which is the power data Dp during the monitoring period PT3, for each of the multiple power generation units 78 that are connected to the same PCS 8 as the target power generation unit 78 and whose flags F1, F2, and F3 are all off (step S82).

[0221] Next, the detection device 101 calculates the median value MEDtm for each measurement time tm in the monitoring period PT3 using the multiple acquired reference power data DpR (step S83).

[0222] Next, the detection device 101 calculates an output ratio R3tm, which is a value obtained by dividing the actual power value PMtm included in the power data Dp of the target power generation unit 78 by the median value MEDtm, for each measurement time tm in the monitoring period PT3 (step S84).

[0223] Next, the detection device 101 classifies all the calculated output ratios R3tm into output ratios R3Aam, which are the output ratios R3tm in the numerical range RgA corresponding to the measurement time tm in the morning, output ratios R3Apm, which are the output ratios R3tm in the numerical range RgA corresponding to the measurement time tm in the afternoon, output ratios R3Bam, which are the output ratios R3tm in the numerical range RgB corresponding to the measurement time tm in the morning, and output ratios R3Bpm, which are the output ratios R3tm in the numerical range RgB corresponding to the measurement time tm in the afternoon.Then, the detection device 101 extracts 100 output ratios R3B1 and 100 output ratios R3B2 from which abnormal values ​​and outliers have been removed (step S85).

[0224] Next, if the increase rate IR3 of the output ratios R3B1, R3B2 during the period from the day the flag F3 of the target power generation unit 78 was turned on to the current day is greater than or equal to the threshold value Thx6 (NO in step S86), the detection device 101 switches the flag F3 of the target power generation unit 78 from on to off (step S87) and terminates the post-processing for the target power generation unit 78.

[0225] On the other hand, if the increase rate IR3 is less than the threshold value Thx6 (YES in step S86), the detection device 101 keeps the flag F3 of the target power generation unit 78 on (step S88) and ends the post-processing for that power generation unit 78.

[0226] In the detection device 101 according to the embodiment of the present disclosure, the processing unit 32 is configured to perform the determination process based on the power data Dp, but this is not limited to the above. The processing unit 32 may be configured to perform the determination process based on the current data Dc instead of the power data Dp.

[0227] In addition, in the detection device 101 according to the embodiment of the present disclosure, the storage unit 33 is configured to store the threshold value Th set based on the output characteristics of the normal string and the output characteristics of the reduced output string, but this is not limited to this. For example, at least one of the lower threshold value ThL and the upper threshold value ThH may be a value that is not based on the output characteristics of the normal string and the output characteristics of the reduced output string.

[0228] In addition, in the detection device 101 according to the embodiment of the present disclosure, the storage unit 33 is configured to store the lower threshold value ThL that is set based on the output power Wa1 of the normal string and the output power Wa2 of the output-reduced string, but this is not limiting. For example, the lower threshold value ThL may be a value that is not based on the output powers Wa1 and Wa2.

[0229] In addition, in the detection device 101 according to the embodiment of the present disclosure, the processing unit 32 is configured to perform an extraction process to extract actual power values ​​PMtm for a time period excluding an excluded time period from among actual power values ​​PMtm included in the power data Dp of the target power generation unit 78 and the reference power generation unit 78, but this is not limited to this. The processing unit 32 may be configured not to perform the extraction process. In this case, the processing unit 32 generates the estimation formula FM using all actual power values ​​PMtm for the calculation target period PT1 acquired from the storage unit 33.

[0230] Furthermore, in the detection device 101 according to the embodiment of the present disclosure, the processing unit 32 is configured to generate the estimation formula FM using the actual power value PMtm in the calculation target period PT1 that is earlier than the determination target period PT2, but this is not limited to the above. The processing unit 32 may be configured to generate the estimation formula FM using the actual power value PMtm in a past period that partially overlaps with the determination target period PT2, or may be configured to generate the estimation formula FM using the actual power value PMtm in the determination target period PT2.

[0231] In addition, in the detection device 101 according to the embodiment of the present disclosure, the processing unit 32 is configured to perform post-processing, but this is not limited thereto. The processing unit 32 may be configured not to perform post-processing. Furthermore, the processing unit 32 may be configured to perform post-processing for flag F1, but not perform post-processing for flags F2 and F3. When the processing unit 32 does not perform post-processing for flag F2 and F3, in the post-processing for flag F1, the processing unit 32 determines that an abnormality has occurred in the target power generation unit 78 when the elapsed time from the day when the flag F1 of the target power generation unit 78 was turned on by maintaining the flag F1 of the target power generation unit 78 on reaches one month.

[0232] In addition, in the detection device 101 according to the embodiment of the present disclosure, the processing unit 32 is configured to perform the determination process 1, the determination process 2, the determination process 3, and the determination process 4, but this is not limited to this. The processing unit 32 may be configured not to perform some of the determination processes 1, 2, 3, and 4.

[0233] In addition, in the detection device 101 according to the embodiment of the present disclosure, the processing unit 32 is configured to perform pre-processing and main processing while not performing post-processing when the flags F1, F2, and F3 of the target power generation unit 78 are all off, but this is not limited to the above. The processing unit 32 may be configured to perform post-processing in addition to pre-processing and main processing when the flags F1, F2, and F3 of the target power generation unit 78 are all off.

[0234] In addition, in the detection device 101 according to the embodiment of the present disclosure, the processing unit 32 is configured to perform post-processing while not performing pre-processing and main processing when any one of the flags F1, F2, and F3 of the target power generation unit 78 is on, but this is not limited thereto. The processing unit 32 may be configured to perform pre-processing and main processing in addition to post-processing when any one of the flags F1, F2, and F3 of the target power generation unit 78 is on. This makes it possible to determine a further output reduction equivalent to the power generation amount of the cluster 92 due to a short circuit of a bypass diode 93 other than the bypass diode 93 in the target power generation unit 78 in which an output reduction equivalent to the power generation amount of the cluster 92 has already occurred due to a short circuit of the bypass diode 93. More specifically, when the processing unit 32 determines that a further output reduction due to a short circuit of the other bypass diode 93 has occurred in the target power generation unit 78, it turns on another flag F1 in the target power generation unit 78 and performs post-processing for each of the multiple flags F1. This makes it possible to detect a state in which a short circuit of multiple bypass diodes 93 is maintained for a long period of time as an abnormality for each bypass diode 93 individually. For example, a state in which the bypass diodes 93, the number of which is 10% or less of the number of clusters 92 in the target power generation unit 78, remain short-circuited for a long period of time can be detected as an individual abnormality.

[0235] Alternatively, the processing unit 32 may be configured to, in post-processing of the flag F3 of the target power generation unit 78, when the period of time that has elapsed since the day on which the flag F3 of the target power generation unit 78 was turned on reaches a predetermined value, switch the flag F3 off and perform notification processing to notify the administrator of the solar power generation system 401 of a request to deal with the abnormality and the start of detection of a new abnormality, etc.

[0236] <Modifications of post-processing> The processing unit 32 may perform a modification process described below instead of or in addition to the post-processing described above. As the modification process, the processing unit 32 performs a detection process for detecting an abnormality in the target power generation unit 78 that has been determined to have an output reduction in the determination process of the main process.

[0237] (1) Transformation process for flag F1 The processing unit 32 judges whether or not to validate the judgment result that the output of the target power generation unit 78 has decreased in the judgment process, based on the length of the period during which the proportion of the actual power values ​​PMtm indicating an output reduction equivalent to the power generation amount of the cluster 92 in the target power generation unit 78, among the multiple actual power values ​​PMtm in the classification period PT4 including the measurement time tm of the actual power values ​​PMtm used to calculate the evaluation value RV, continues to be equal to or greater than a predetermined value.

[0238] More specifically, the processing unit 32 acquires the power data Dp of the target power generation unit 78 for the classification period PT4 from the storage unit 33. The classification period PT4 is, for example, the period from one month before the day when the flag F1 of the target power generation unit 78 was turned on to the day of the classification.

[0239] The processing unit 32 uses a machine learning method such as kNN to classify the actual power values ​​PMtm included in the acquired power data Dp for the classification period PT4 into an actual power value PMtmD1, an actual power value PMtmD2, an actual power value PMtmD3 or an actual power value PMtmD4.

[0240] Here, the actual power value PMtmD1 is the actual power value PMtm indicating that the output of the target power generation unit 78 is showing a tendency to recover. The actual power value PMtmD2 is the actual power value PMtm indicating that the output of the target power generation unit 78 is normal. The actual power value PMtmD3 is the actual power value PMtm indicating that the target power generation unit 78 has experienced an output reduction corresponding to the amount of power generated by the cluster 92. The actual power value PMtmD4 is the actual power value PMtm indicating that the target power generation unit 78 has experienced an output reduction due to the influence of shading.

[0241] The processing unit 32 calculates, for each day in the classification period PT4, a ratio Rx1 of the number of actual power values ​​PMtmD3 to the number of actual power values ​​PMtm of the target power generation unit 78. The processing unit 32 compares the calculated ratio Rx1 with a predetermined threshold value Thy.

[0242] If the rate Rx1 falls below the threshold Thy before the period during which the rate Rx1 remains equal to or greater than the threshold Thy reaches one month, the processing unit 32 switches the flag F1 of the target power generation unit 78 from on to off.

[0243] On the other hand, if the period during which the ratio Rx1 remains greater than or equal to the threshold value Thy reaches one month, the processing unit 32 determines that there is a possibility that an abnormality has occurred in the target power generation unit 78, and switches off the flag F1 of the target power generation unit 78 and switches on the flag F2 of the target power generation unit 78.

[0244] (2) Transformation process for flag F2 The processing unit 32 acquires from the storage unit 33 the power data Dp of the target power generation unit 78 for the classification period PT4.

[0245] The processing unit 32 uses a machine learning method such as kNN to classify the actual power values ​​PMtm included in the power data Dp in the classification period PT4 into an actual power value PMtmD1, an actual power value PMtmD2, an actual power value PMtmD3 or an actual power value PMtmD4.

[0246] The processing unit 32 calculates, for each day in the classification period PT4, a ratio Rx2 of the number of actual power values ​​PMtmD3 to the number of actual power values ​​PMtm of the target power generation unit 78. The processing unit 32 compares the calculated ratio Rx2 with a threshold value Thy.

[0247] If the rate Rx2 falls below the threshold Thy before the period during which the rate Rx2 remains equal to or greater than the threshold Thy reaches one month, the processing unit 32 switches the flag F2 of the target power generation unit 78 from on to off.

[0248] On the other hand, if the period during which the ratio Rx2 remains greater than or equal to the threshold value Thy reaches one month, the processing unit 32 determines that an abnormality has occurred in the target power generation unit 78, switches off the flag F2 of the target power generation unit 78, and switches on the flag F3 of the target power generation unit 78.

[0249] (3) Transformation process for flag F3 The processing unit 32 acquires from the storage unit 33 the power data Dp of the target power generation unit 78 for the classification period PT4.

[0250] The processing unit 32 uses a machine learning method such as kNN to classify the actual power values ​​PMtm included in the power data Dp in the classification period PT4 into an actual power value PMtmD1, an actual power value PMtmD2, an actual power value PMtmD3 or an actual power value PMtmD4.

[0251] The processing unit 32 calculates, for each day in the classification period PT4, a ratio Rx3 of the number of actual power values ​​PMtmD3 to the number of actual power values ​​PMtm of the target power generation unit 78. The processing unit 32 compares the calculated ratio Rx3 with a threshold value Thy.

[0252] If the rate Rx3 falls below the threshold Thy before the period during which the rate Rx3 remains equal to or greater than the threshold Thy reaches one month, the processing unit 32 switches the flag F3 of the target power generation unit 78 from on to off.

[0253] The above-described embodiments should be considered as illustrative and not restrictive in all respects. The scope of the present invention is defined by the claims, not by the above description, and is intended to include all modifications within the meaning and scope of the claims.

[0254] Each process (each function) of the above-mentioned embodiment is realized by a processing circuit (circuitry) including one or more processors. The above-mentioned processing circuit may be composed of an integrated circuit or the like in which one or more memories, various analog circuits, and various digital circuits are combined in addition to the above-mentioned one or more processors. The above-mentioned one or more memories store programs (instructions) that cause the above-mentioned one or more processors to execute each of the above-mentioned processes. The above-mentioned one or more processors may execute each of the above-mentioned processes according to the programs read from the above-mentioned one or more memories, or may execute each of the above-mentioned processes according to a logic circuit designed in advance to execute each of the above-mentioned processes. The above-mentioned processors may be various processors suitable for computer control, such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a DSP (Digital Signal Processor), an FPGA (Field Programmable Gate Array), and an ASIC (Application Specific Integrated Circuit). The above-mentioned physically separated processors may execute each of the above-mentioned processes in cooperation with each other. For example, the processors mounted on each of a plurality of physically separated computers may cooperate with each other via a network such as a LAN (Local Area Network), a WAN (Wide Area Network), the Internet, etc. The program may be installed in the memory from an external server device or the like via the network, or may be distributed in a state stored in a recording medium such as a CD-ROM (Compact Disc Read Only Memory), a DVD-ROM (Digital Versatile Disc Read Only Memory), or a semiconductor memory, and installed in the memory from the recording medium.

[0255] The above description includes the following additional features. [Appendix 1] A determination device for use in a solar power generation system having a plurality of strings including solar cell panels, a threshold acquisition unit that acquires a threshold that is set based on the magnitude of the output of a target portion of the solar panel; a performance value acquisition unit for acquiring a performance value of the output of each of the strings; a predicted value calculation unit that calculates a predicted value of an output of the target string at a target time based on the actual value of an output of a target string among the plurality of strings and the actual value of an output of a reference string among the plurality of strings, both of which are acquired by the actual value acquisition unit; a feature calculation unit that calculates a feature that is a ratio between the predicted value calculated by the predicted value calculation unit and the actual value of the output of the target string at the target time; a determination unit that performs a determination process to determine an output reduction corresponding to an amount of power generation of the target portion in the target string based on a comparison result between an evaluation value based on the feature amount calculated by the feature amount calculation unit and the threshold value acquired by the threshold acquisition unit, the solar panel includes a plurality of clusters having series circuits of solar cells; A determination device, wherein the target portion is one of the clusters.

[0256] [Appendix 2] A determination device for use in a solar power generation system having a plurality of strings including solar cell panels, A processing circuit is provided, The processing circuitry includes: obtaining a threshold value that is set based on the magnitude of the output of a target portion of the solar panel; obtaining an actual value of the output of each of said strings; calculating a predicted value of an output of the target string at a target time based on the acquired actual value of an output of a target string among the plurality of strings and the acquired actual value of an output of a reference string among the plurality of strings; Calculating a feature amount that is a ratio between the calculated predicted value and the actual value of the output of the target string at the target time; A determination device that performs a determination process to determine an output reduction in the target string that corresponds to the amount of power generated by the target portion based on the result of comparing an evaluation value based on the calculated feature amount with the acquired threshold value. [Explanation of symbols]

[0257] 1 output line 1p positive output line 1n Negative output line 2,4,5 Aggregation Line 3. Internal Lines 5p positive side aggregate line 5n negative side aggregation line 6. Cubicle 7,72,73,77 Copper bars 8 PCS 9 Power conversion section 11 Acquisition Department 14 Communications Department 16 Current Sensor 17 Voltage Sensor 26 Power line 26p positive power line 26n Negative power line 31 Acquisition Section (Actual Value Acquisition Section) 32 Processing unit (threshold acquisition unit, predicted value calculation unit, feature amount calculation unit, judgment unit) 33 Storage section 46 Conductor 60 Current collector unit 71 Current collector box 75 Solar Cell Unit 76 Junction box 77p positive copper bar 77n negative copper bar 78 Power Generation Division 79 Solar Panels 80 PCS units 91 Solar Cell 92, 92A, 92B, 92C Cluster (target part) 93, 93A, 93B, 93C Bypass diodes (bypass circuits) 101 Detection device 111 Monitoring equipment 151 Collection Device 301 Power Generation Status Management System 401 Photovoltaic Power Generation System N1, N2, N3, N4, N5, N6 nodes Cpv1, Cpv2 PV curve Vdif Difference Vp1 voltage value (first voltage value) Vp2 voltage value (second voltage value) Vp3 voltage value (third voltage value) Wa1, Wa2, Wb1, Wb2 Output power

Claims

1. A determination device for use in a solar power generation system having a plurality of strings including solar cell panels, a threshold acquisition unit that acquires a threshold that is set based on the magnitude of the output of a target portion of the solar panel; a performance value acquisition unit for acquiring a performance value of the output of each of the strings; a predicted value calculation unit that calculates a predicted value of an output of the target string at a target time based on the actual value of an output of a target string among the plurality of strings and the actual value of an output of a reference string among the plurality of strings, both of which are acquired by the actual value acquisition unit; a feature calculation unit that calculates a feature that is a ratio between the predicted value calculated by the predicted value calculation unit and the actual value of the output of the target string at the target time; a determination unit that performs a determination process to determine an output reduction in the target string that corresponds to the power generation amount of the target portion based on a comparison result between an evaluation value based on the feature calculated by the feature calculation unit and the threshold acquired by the threshold acquisition unit.

2. the target portion includes a bypass circuit; The determination device according to claim 1, wherein the threshold value is set based on the output characteristics of a normal string, which is the string including the solar panel in which the output of the target portion is normal, and the output characteristics of a reduced output string, which is the string including the solar panel in which the output of the target portion is reduced due to a short circuit in the bypass circuit.

3. the threshold is set based on an output of the normal string at a third voltage value and an output of the degraded string at the third voltage value; 3. The determination device of claim 2, wherein the third voltage value is greater than a first voltage value corresponding to a peak power in the power-voltage characteristic of the normal string and is smaller than a value obtained by adding an absolute value of a difference between the first voltage value and a second voltage value corresponding to a peak power in the power-voltage characteristic of the reduced output string to the first voltage value.

4. The determination device of claim 1, wherein the predicted value calculation unit calculates the predicted value by excluding the actual values ​​for time periods in which at least one of the actual value of the output of the target string, the actual value of the output of the reference string, and the ratio of the actual value of the output of the target string to the actual value of the output of the reference string satisfies a predetermined condition.

5. the determination unit performs the determination process using the evaluation value based on the feature amounts at a plurality of target times in a determination target period; The determination device according to claim 1 , wherein the predicted value calculation unit calculates the predicted value based on the actual value from a time period prior to the determination target time period.

6. The determination device according to any one of claims 1 to 5, wherein the determination unit determines whether or not to validate a determination result that the output of the target string has decreased in the determination process based on an output ratio that is a ratio between a representative value of the actual values ​​of the outputs of the multiple strings and the actual value of the output of the target string.

7. 7. The determination device according to claim 6, wherein the determination unit invalidates the determination result when a difference between a plurality of the output ratios corresponding to a plurality of measurement timings of the performance value and an output ratio within a predetermined range among the plurality of output ratios is equal to or greater than a predetermined value.

8. The determination device according to claim 6 , wherein the determination unit invalidates the determination result when a rate of decrease in the output ratio in a period before the target time is equal to or greater than a predetermined value.

9. The determination device according to claim 6 , wherein the determination unit invalidates the determination result when a rate of decrease in the output ratio in the period after the target time is equal to or greater than a predetermined value.

10. 7. The determination device according to claim 6, wherein the determination unit invalidates the determination result when an increase rate of the output ratio within a predetermined range among a plurality of the output ratios corresponding to a plurality of measurement timings in a period after the target time is equal to or greater than a predetermined value.

11. The determination device according to any one of claims 1 to 5, wherein the determination unit determines whether to validate a determination result that the output of the target string has decreased in the determination process based on the length of a period during which a ratio of the actual values ​​indicating an output decrease equivalent to the amount of power generated in the target part of the target string, among the multiple actual values ​​during a period including the target time, continues to be equal to or greater than a predetermined value.

12. A determination method for a determination device used in a solar power generation system including a plurality of strings including solar cell panels, comprising: Obtaining a threshold value that is set based on the magnitude of the output of a target portion of the solar panel; obtaining a performance value of the output of each of said strings; calculating a predicted value of an output of the target string at a target time based on the acquired actual value of an output of a target string among the plurality of strings and the acquired actual value of an output of a reference string among the plurality of strings; calculating a feature value that is a ratio between the calculated predicted value and the actual value of the output of the target string at the target time; The method includes a step of performing a determination process to determine an output reduction in the target string that corresponds to the amount of power generated by the target portion based on a comparison result between an evaluation value based on the calculated feature and the acquired threshold value.

13. A determination program for use in a determination device for a solar power generation system having a plurality of strings including solar cell panels, comprising: Computer, a threshold acquisition unit that acquires a threshold that is set based on the magnitude of the output of a target portion of the solar panel; a performance value acquisition unit for acquiring a performance value of the output of each of the strings; a predicted value calculation unit that calculates a predicted value of an output of the target string at a target time based on the actual value of an output of a target string among the plurality of strings and the actual value of an output of a reference string among the plurality of strings, both of which are acquired by the actual value acquisition unit; a feature calculation unit that calculates a feature that is a ratio between the predicted value calculated by the predicted value calculation unit and the actual value of the output of the target string at the target time; a determination unit that performs a determination process to determine an output reduction corresponding to the power generation amount of the target portion in the target string based on a comparison result between an evaluation value based on the feature amount calculated by the feature amount calculation unit and the threshold value acquired by the threshold acquisition unit; A judgment program to function as a.

Images