JP6555152B2 - Solar cell module system - Google Patents

Description

  The present invention relates to a solar cell module system.

  The solar cell system appropriately adjusts the amount of power generation while measuring the output generated by the solar cell (solar cell) with a measurement sensor such as a current sensor and / or a voltage sensor.

  However, characteristics of the current sensor and / or voltage sensor used for adjusting the power generation amount may be shifted due to manufacturing variations, aging deterioration, temperature characteristics, and the like.

  Among the above-mentioned deviations, manufacturing deviations can be eliminated by storing the initial characteristics in a nonvolatile memory and correcting in advance, and applying the variation amount to the calculation during dynamic learning. Further, since aging degradation is a gradual change, the deviation can be appropriately corrected by detecting the amount of change for each long-term operation, for example, at the time of starting to start the power generation operation.

  However, since the temperature characteristic is a high time responsive shift that changes in a short period, it is desirable to eliminate the temperature characteristic shift periodically during measurement of the power generation operation in which the temperature changes.

  Here, in order to correct | amend according to temperature in the sensor used for the said electric power generation amount adjustment, it is necessary to stop control of a solar cell module temporarily. For example, in the case of a voltage sensor, a state in which the current is temporarily made zero is used as a reference, and the state of deviation from the reference of the measured voltage (output value) is calculated from the deviation from the reference.

  For example, in the solar cell module of Patent Document 1, when the measured temperature becomes a predetermined temperature difference from the start temperature, control for power generation is temporarily stopped and the output value measured by the measurement sensor is corrected each time. Therefore, the learning operation for calculating the deviation from the reference is controlled.

Japanese Patent Laying-Open No. 2015-136233

  When the above control is used, every time a predetermined temperature difference is reached, a learning operation for calculating a deviation from the reference value of the measurement value is performed in order to suspend the control for power generation and correct the output value. In the learning period for the correction, the power generation amount of the solar cell module is reduced.

  In view of the above circumstances, an object of the present invention is to provide a solar cell module system that can suppress a decrease in the amount of power generation.

In order to solve the above problems, according to one aspect of the present invention, a solar panel in which a plurality of solar cells are arranged; a measurement sensor that measures an output value of a current or voltage output from the solar panel; A temperature sensor for measuring temperature; a correction unit for correcting an output value measured by the measurement sensor based on a temperature measured by the temperature sensor; and for correcting an output value measured by the measurement sensor A sensor characteristic learning unit that performs sensor characteristic learning; a learning temperature point that is a temperature at which the sensor characteristic learning is performed and a learning condition calculation unit that calculates the number of learnings; and a temperature of the measurement sensor measured by the temperature sensor Is a solar cell module system comprising: a temperature determination unit that determines whether or not the learning temperature point has been reached;
In the sensor characteristic learning, by temporarily disconnecting the measurement sensor and the solar panel, a current value or a voltage value generated by a constant current source or a constant voltage source provided in the vicinity of the measurement sensor, and Comparing the generated current value or voltage value with the output value measured by flowing or applied to the measurement sensor to grasp the deviation,
The correction unit corrects an output value measured by the measurement sensor based on the deviation,
When the sensor characteristic learning unit performs the startup sensor characteristic learning at the startup ambient temperature Tenv, the learning condition calculation unit needs to re-learn the temperature increase range ΔTsys during the system operation and the measurement sensor. According to the temperature difference ΔTstdth, the learning number Nstd excluding start-up and the optimum learning temperature point Tstd (i) that becomes the center temperature of sensor characteristic learning during operation;
Calculate the learning count Nstd = (ΔTsys / ΔTstdth) by rounding it down to the nearest lower positive number,
Set the optimal learning temperature point “Tstd (i) = Tenv + ΔTsys / 2−ΔTstdth / 2 × (Nstd−1) + ΔTstdth × i” (i = 0, 1, 2,... (Nstd−1))
When the temperature determination unit determines that the temperature measured by the temperature sensor has reached the optimal learning temperature point Tstd (i), the sensor characteristic learning unit detects the optimal learning temperature point Tstd (i). Sensor characteristic learning is performed during operation.

  By setting the optimal learning temperature point according to the above settings, the temperature learning width range during solar cell module system operation can be properly covered with the optimal learning temperature point ± temperature difference required for re-learning of the measurement sensor. Can do.

  Therefore, it is possible to reduce the number of executions of the sensor characteristic learning that accompanies the temporary stop of the power generation of the solar cell, and to suppress a decrease in the amount of power generation.

  According to one aspect, in the solar cell module system, by setting the optimal learning temperature point for performing sensor characteristic learning of the measurement sensor, it is possible to appropriately cover the range of the temperature increase range during system operation, so that the solar It is possible to reduce the number of times of sensor characteristic learning that is accompanied by temporary stoppage of power generation of the panel, and to suppress a decrease in the amount of power generation.

The block diagram of control of a 1st embodiment of the present invention. The graph which shows the sensor characteristic learning point in a comparative example, and the operating condition of system operation | movement. The flowchart of 1st Embodiment of this invention. The example of selection of the learning point in a comparative example and 1st Embodiment. The block diagram of control of 2nd Embodiment of this invention. The flowchart of 2nd Embodiment. Example of equal allocation of relearning points. The graph which shows reuse of a learning result and the frequency | count of learning in a comparative example and this invention.

  Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings.

[First Embodiment]
FIG. 1 shows a block diagram of control according to the first embodiment of the present invention. The power generation system 1 according to the present invention is a solar cell module system that performs solar power generation.

  The power generation system (solar cell module system) 1 includes a solar panel 10, a current sensor 20, a temperature sensor 30, a measurement value control unit 40, a power generation control unit 50, and a power conversion unit 60. The power conversion unit 60 supplies the generated power to the supply destination 70 via the power system.

  The solar panel 10 has a string configuration (11) in which a plurality of solar cells (solar cells) are connected in series, and a bypass diode is appropriately provided for each of a plurality of units.

  The current sensor 20 measures the output current of the solar panel 10. Here, a current sensor is cited as an example of the measurement sensor, but the measurement sensor used in the present invention is, for example, a current sensor, a voltage sensor, or the like, and the measurement result changes depending on the temperature (deviates from the reference value). Any sensor can be used.

  The temperature sensor 30 is provided in the vicinity of the current sensor 20 and measures the temperature of the current sensor 20.

  The measurement value control unit 40 corrects the output value of the current sensor 20 whose measurement characteristics change according to the temperature measured by the temperature sensor 30, and outputs the correction value to the power generation control unit 50 and the power conversion unit 60. The measurement value control unit (correction unit) 40 includes a learning condition calculation unit 41, a learning temperature storage memory 42, a learning temperature determination unit 43, a sensor characteristic learning unit 44, and a measurement value correction unit 45.

  A constant current source 21 is provided in the vicinity of the current sensor 20 and is used for learning of sensor characteristics depending on the temperature of the current sensor 20 in the sensor characteristic learning unit 44 of the measurement value control unit 40. When the measurement sensor is a voltage sensor, a constant voltage source is provided and a predetermined voltage is applied.

  The power generation control unit 50 manages the power generation status of the solar panel 10 based on the output value (output current) corrected by the measurement value correction unit 45 of the measurement value control unit 40.

  The power conversion unit 60 is, for example, a DCDC converter or the like, and the load, applied voltage / current, and the like are adjusted by feedback control according to an instruction from the power generation control unit 50 so that the amount of power in the solar panel 10 is optimized. Be made.

  This system is a solar system that performs power generation control using, for example, MPPT (Maximum Power Point Tracking) control or PWM control.

  Therefore, in this system, the power generation control unit 50 monitors the output value (corrected output value) of the current and / or voltage output from the solar panel 10 measured by the current sensor (or / and voltage sensor). In accordance with the monitoring result, the power generation control unit 50 adjusts the power conversion unit 60 as appropriate so that the power generation amount extracted from the solar panel 10 is as large as possible.

  Here, although solar panels generally generate power by absorbing light, a portion of the absorbed light changes to heat instead of power, and the surface temperature rises.

  The current sensor and / or voltage sensor provided in the vicinity of the solar panel 10 and used for the adjustment of the power generation amount may shift in characteristics due to temperature characteristics in addition to manufacturing variations and aging degradation. .

  Table 1 shows the characteristics of various deviations. In Table 1, ○ indicates that correction is possible, and among the deviations shown in Table 1, when correction is possible by a plurality of methods, correction is performed as few times as possible (indicated by ◎). .

表１において、（１）製造ばらつきは、初期特性を不揮発性メモリに記憶して予め補正し、動的学習の際に、ばらつき量を演算に適用することで、ズレを解消することができる。そのため、製造の際に、１回だけ補正すれば済む。

In Table 1, (1) manufacturing variations can be eliminated by storing initial characteristics in a nonvolatile memory and correcting in advance, and applying the variation amount to the calculation during dynamic learning. Therefore, only one correction is required during manufacturing.

  In addition, (2) aging deterioration is sufficient if a characteristic is detected once every long period of time, for example, at the time of system startup, for each predetermined period.

  However, (3) Since the temperature characteristics are highly time-dependent deviations that change in a short period, the temperature characteristics deviations are periodically grasped (calculated) based on the characteristic map during use. It is necessary to dynamically correct every short time.

  In the present invention, in the “sensor characteristic learning operation” of the output value, the control is temporarily stopped, a reference state is created, and the deviation of the measured value is calculated based on the deviation from the reference. means. In the case of a voltage sensor, a state in which the current is temporarily made zero is created and used as a reference, and the degree of deviation of the generated voltage is calculated from the deviation from the reference.

  Specifically, in the sensor characteristic learning operation, the connection of the current sensor 20 with the solar panel 10 is disconnected and temporarily interrupted, and the constant current source 21 provided near the current sensor 20 is used. Current is passed to 20.

  First, the sensor characteristic learning unit 44 compares the current value generated by the constant current source 21 with the output value measured when the generated current value flows to the current sensor 20 at the ambient temperature Tenv at the time of startup. Conduct sensor characteristics learning at start-up to grasp the error (error).

  Further, when the temperature of the current sensor 20 measured by the temperature sensor 30 changes to learning temperature points Tstd (0), Tstd (1), Tstd (2), which will be described later, the sensor characteristic learning unit 44 The current value generated by the constant current source 21 and the current value measured when the generated current value flows to the current sensor 20 are compared to grasp and record the deviation.

  Here, in the temperature rise width ΔTsys which fluctuates in the solar cell module system 1, the above deviation is calculated and stored when the temperature rises for the first time, and then the temperature of the current sensor 20 in the system temperature rise width ΔTsys is calculated. If known, the deviation (sensor characteristics) of the current sensor 20 can be known at that temperature. Therefore, the deviation can be recognized and corrected as appropriate without performing comparison again.

  Note that the stored temperature differences of learning temperature points Tstd (0), Tstd (1), Tstd (2), and Tstd (3) are within the guaranteed operating temperature range + ΔTstdth (re-learning necessary temperature difference). is there. Therefore, in this embodiment, when the measured temperature reaches the closest learning temperature point Tstd (i), a learning operation for correcting the compared deviation may be performed.

  The comparison operation at the ambient temperature Tenv at the start of system operation is called the sensor learning operation at startup, and the comparison operation at Tstd (0), Tstd (1), Tstd (2), is called the active sensor learning operation. . The sensor learning operation at startup and the sensor learning operation during operation are collectively referred to as a learning operation.

<How to shift>
FIG. 2 shows an example of sensor characteristics, and a graph showing sensor characteristics learning points and system operation statuses in a comparative example.

  In this specification, T indicates the temperature of the current sensor (measurement sensor) 20 measured by the temperature sensor 30. Tenv is the ambient temperature around the measurement sensor (Environment) and means the temperature at the start of operation. ΔTsys indicates a temperature increase width during operation of the solar cell module system 1 (hereinafter referred to as a temperature increase width during system operation). ΔTstdth indicates a temperature difference required for re-learning of sensor characteristics. That is, ΔTstdth means an operation guaranteed temperature range when the temperature of the measurement sensor is changed.

  Assuming that the sensor characteristic can be expressed by a linear expression (y = ax + b: x is an actual value, y is a recognized value) as a method of deviation of the output value of the current sensor 20 depending on temperature, a is Gain and b are offset.

  In order to correct the offset b in the above equation, one set of correct values and measurement values is required (for example, the y-intercept when x = 0). Further, in order to correct the gain a, two sets of correct values and measured values are required for adjusting the slope of the linear expression.

  Here, when dynamic correction is necessary, in order to obtain the measured value of (X, Y) on-board where the solar cell module system 1 is mounted on another system, normal control is performed. It is necessary to stop and implement the measurement mode. That is, the control unit 40 temporarily stops the power generation of the solar panel 10 during the period during which the sensor characteristic learning during operation for correcting the output value measured by the measurement sensor 20 is performed (period of the measurement mode).

  However, in the measurement mode, since power generation is temporarily stopped and normal control cannot be performed, the original function is impaired, and the power generation amount is reduced accordingly. Therefore, it is preferable that the number of measurement modes is smaller.

  FIG. 3 shows a flowchart of the measurement mode according to the first embodiment of the present invention. In the flow of the present invention shown in FIG. 3, the control of S2 and S4 is the same as that of the comparative example shown in FIG. 2, except that the steps S1 and S3 are provided.

  First, when the system operation is started, before starting the power generation operation, during the control stop period (in the measurement mode), the startup sensor characteristic learning operation is performed at the ambient temperature Tenv that is the start temperature (S2). In parallel, a learning temperature point Tstd (i) is calculated (S1).

  As described above, in S2, by comparing the current value generated by the constant current source 21 with the current value measured when the generated current flows to the current sensor 20 at the ambient temperature Tenv at the time of startup. The start-up sensor characteristic learning operation for grasping the deviation of the current sensor 20 at the start temperature is performed.

  A method in which the learning condition calculation unit 41 of the measurement value control unit 40 calculates the learning temperature point Tstd (i) and the learning number Nstd in S1 will be described in detail.

  First, the temperature increase width ΔTsys during system operation is acquired by setting the power generation control unit 50 and the like, and at least the re-learning necessary temperature difference ΔTstdth of the current sensor 20 that eliminates the manufacturing variation is acquired from, for example, the learning temperature storage memory 42 To do.

  Then, as S1-1, the learning condition calculation unit 41 uses the following equation to calculate the learning score (the learning score excluding the start-up time) Nstd using the temperature rise width ΔTsys during system operation and the temperature difference ΔTstdth necessary for the sensor to be re-learned. calculate. The learning point number Nstd means the number of corrections excluding start-up when performing sensor characteristic learning for correcting the output value measured by the measurement sensor based on the measured temperature.

Nstd = floor (ΔTsys / ΔTstdth) (Formula 1)
In the above equation 1, (ΔTsys / ΔTstdth) is rounded down to the nearest lower positive number by the floor function.

Then, a learning temperature point is calculated as S1-2. The learning temperature point means the center temperature of sensor characteristic learning during operation. Where I = 0, 1, 2, 3,... (Nstd-1)
Tstd (i) = Tenv + ΔTsys / 2−ΔTstdth / 2 × (Nstd−1) + ΔTstdth × i (Expression 2)
Is calculated by

For example, when the temperature increase width ΔTsys during system operation is smaller than twice the temperature difference ΔTstdth necessary for re-learning of the sensor (ΔTsys ≦ 2 × ΔTstdth), Nstd = 1 and I = 0. Therefore
Tstd (i) = Tenv + ΔTsys / 2−ΔTstdth / 2 × (1-1) + ΔTstdth × 0
= Tenv + ΔTsys / 2.
Thus, the calculated learning temperature point Tstd (i) is stored in the learning temperature storage memory 42.

  In parallel with the calculation operation of S1, in S2, the sensor characteristic learning unit 44 executes the startup sensor characteristic learning operation at the start temperature (ambient temperature) Tenv.

  Thereafter, when the system operation is started, the current value (output current) measured by the current sensor 20 is used for power generation control of the solar panel 10 whose temperature increases with power generation, and the temperature of the current sensor 20 that increases. The measurement temperature T is measured by the temperature sensor 30.

  In S3, it is determined to which learning temperature point Tstd (i) the measured temperature T is close (learning temperature determination step).

In other words, it is determined which temperature range the measured temperature T matches.
T ≒ Tstd (i) I = 0, 1, 2, 3, ... (Nstd-1)
That is, it is determined in which range the measured temperature T is Tstd (0) ± ΔTstdth, Tstd (1) ± ΔTstdth, Tstd (2) ± ΔTstdth, Tstd (Nstd−1) ± ΔTstdth To do.

  As shown on the right side of FIG. 4 to be described later, since the arrow indicating the operation guarantee range overlaps the range of + ΔTstdth, the coincidence of the ranges is determined by the learning temperature determination unit 43 by measuring the temperature T However, it is only necessary to detect (determine) whether or not the learning temperature point Tstd (i) has been reached.

  As described above, it is preferable that the coincidence determination has a width in accordance with the control cycle of the system and the accuracy of the current sensor 20 and the temperature sensor 30 in the determination of S3.

  When the measured temperature T reaches the learning temperature point Tstd (i), the measured value correcting unit 45 outputs the output measured by the current sensor 20 while causing the power generation control unit 50 to temporarily stop the power generation of the solar panel 10. Sensor characteristic learning during operation for correcting the value is performed (S4).

  Specifically, when the measured temperature T reaches the learning temperature point Tstd (0), Tstd (1), Tstd (2), Tstd (3), the current value generated by the constant current source 21 and the generated current Is compared with the current value measured by flowing to the current sensor 20, and sensor characteristic learning during operation is performed to grasp the deviation. Based on the deviation at this time, the measurement value correction unit 45 corrects the current value (output value from the current sensor 20).

  FIG. 4 shows a selection example of learning points in the comparative example and the first embodiment. FIG. 4 also shows the correlation between the number of learning points, the number of re-learning times, and the number of learning times in the comparative example and the first embodiment. In FIG. 4, the case where the operating range is (ΔTstdth <ΔTsys <2 × ΔTstdth) will be considered. In FIG. 4, ΔTstdth at the time of temperature rise corresponds to the upper half of the arrow.

  In the comparative example (left side) of FIG. 4, the learning point of the sensor is determined for each predetermined temperature regardless of the operating temperature range of the solar system. Therefore, as shown in FIG. 4, together with the correction at the start time, the time point when the operating temperature rise width ΔTsys exceeds the re-learning necessary temperature difference ΔTstdth is set as a predetermined temperature, and the current sensor 20 is passed every time it passes the predetermined temperature. The learning operation is implemented.

  This learning temperature point “Tenv + ΔTstdth” is on the high temperature side in the temperature rise width ΔTsys during operation of the system and becomes uneven learning, and the lower temperature cannot be covered in the range of ± ΔTstdth centering on “Tenv + ΔTstdth” ( The border of the range).

  In addition, each time it passes through the specified temperature (every time it passes the learning temperature point ("Tenv", "Tenv + ΔTstdth")), the sensor characteristics during operation with the temporary stop of solar panel power generation to perform characteristic learning The number of times of learning has become infinite, and the amount of power generation has decreased accordingly.

  On the other hand, in the control of the present invention (right side in FIG. 4), the correction at the ambient temperature Tenv at the start is performed and “Tenv + ΔTsys / 2” calculated from the above formulas (1) and (2) is corrected as the learning temperature (center temperature). doing.

  By setting an appropriate learning temperature point “Tenv + ΔTsys / 2”, the temperature increase width ΔTsys during all system operations can be covered by + ΔTstdth and −ΔTstdth indicating the difference in temperature required for re-learning on the + side and − side. That is, on the rightmost side of FIG. 4, the dotted line range ΔTsys indicating the upper limit and the lower limit is covered by an arrow (learning temperature range) of the re-learning necessary temperature difference ± ΔTstdth centered on the learning temperature point “Tenv + ΔTsys / 2”. Therefore, the entire range ΔTsys falls within the guaranteed operation range ± ΔTstdth centered on “Tenv + ΔTsys / 2”.

  Therefore, if correction is performed when the temperature rises for the first time, even if the temperature fluctuates afterwards, it is not necessary to perform correction, so re-learning can be omitted. It can be suppressed twice.

  In this way, in the solar cell module system 1, by setting the optimal learning point for performing sensor characteristic learning during operation of the measurement sensor to the temperature, implementation of sensor characteristic learning accompanied by a temporary stop of power generation of the solar panel The number of times can be reduced and the decrease in power generation can be suppressed.

  FIG. 4 shows an example in which the range of the system temperature rise ΔTsys is narrow, the correction sensor learning point Nstd = 1, and all ΔTsys ranges are covered by Tstd (0) ± ΔTstdth. .

  On the other hand, when the temperature decreases when the range of ΔTsys is wide and Nstd = 3 or more, for example, the measured temperature decreases from Tstd (2) → Tstd (1) → Tstd (0), and T = Tstd (0). If the output value of the current sensor 20 is still corrected based on the sensor characteristic (deviation from the current value of the constant current source) at Tstd (2) when it reaches the correction result, the correction result is incorrect.

  Therefore, as described above, a control example when the temperature increase width ΔTsys during system operation is large will be described as a second embodiment.

[Second Embodiment]
FIG. 5 shows a block diagram of the control of the second embodiment of the present invention.

  Even when the current sensor is at the same temperature, the sensor characteristics may slightly change depending on whether the temperature increases from a low temperature or decreases after a high temperature continues.

  In this embodiment, the corrected measurement sensor characteristic (deviation from the current value generated by the constant current source) and the temperature of the measurement sensor at the time of learning for correction (learning temperature point TstdT) are stored in association with each other, The sensor characteristics related to Tstd satisfying TstdT−ΔTstdth <T + ΔTstdth are reused.

  In this control example, as compared with the first embodiment, the correction operation using the learning result is increased by the amount of reuse. However, even when the system temperature increase range ΔTsys is wide, more finely-tuned. It can correspond to. On the other hand, as compared with the comparative example in which learning is performed each time at a predetermined temperature, the number of re-learning (the number of reuse) is about half of the number of learning, so the number of re-learning (the number of reuse) can be reduced.

  The configuration of the present embodiment is different from the first embodiment in that the solar cell module system 2 includes a reuse temperature selection unit 46 and a reuse temperature determination unit 47 in the measurement value control unit 400.

  In the present embodiment, in the storage memory 48, the sensor characteristics (deviation from the current value generated by the constant current source) and the temperature of the current sensor 20 measured by the temperature sensor 30 when it is corrected (learning temperature point). Are stored in association with each other and can be taken out upon reuse.

  In the system of the present embodiment, since there is a resource for storing the learning result, learning can be performed in addition to the first time, and the number of learning can be selectively executed again.

  FIG. 6 shows a flowchart of the second embodiment. In the control of this embodiment shown in FIG. 6, S11, S12, S13, and S14 in FIG. 6 are the same as S1 to S4 in the first embodiment shown in FIG. The control after S15 is different from the control of FIG.

  In S11, as in FIG. 3, the learning temperature point Tstd (i) is calculated, and in S12, the startup learning operation is performed at the start ambient temperature Tenv.

  In S13, similarly to S3, the learning temperature determination unit 43 detects whether the measured temperature T has reached the learning temperature point Tstd (i) through temperature comparison.

  Note that the learning temperature point Tstd (i) is the same as in the first embodiment, and the determination as to which temperature range the measured temperature T matches is T≈Tstd (i) I = 0, 1, 2 , 3, ..., Nstd-1. Since this range overlaps the temperature, as a finer setting, in “optimal setting 1” and “optimal setting 2” of FIG. 8 to be described later, in the determination at the time of learning only once at the first rising temperature. Used.

  Also in this embodiment, it is preferable to give a width to the coincidence determination in accordance with the control cycle of the system and the accuracy of the sensors 20 and 30.

  When the learning temperature point Tstd (i) is reached in S14, the measurement value correction unit 45 corrects the output value measured by the current sensor 20, while causing the power generation control unit 50 to temporarily stop the power generation of the solar panel 10. Conduct sensor characteristics learning.

  Thereafter, in S14, the storage memory 48 stores the learning result at the learning temperature point Tstd (i) (the deviation from the current value generated by the constant current source) in association with the learning temperature point Tstd (i).

  Since this control example is mainly applied when ΔTsys is wide, when the number of corrections Nstd is two or more, S13 to S15 are repeated Nstd times.

  Thereafter, in S16, the reuse temperature selection unit 46 selects a reuse learning point Tstd (e) from Tstd (i) stored in the storage memory 48 in S15. Specifically, the reuse temperature selection unit 46, among the stored Tstd (i), satisfies the learning temperature point TstdT that satisfies “TstdT−ΔTstdth <T <T + ΔTstdth” and the related sensor characteristics (ie, Tstd (e) ( e is a learning characteristic having an even number as a center temperature.

In S17, the reuse temperature determination unit 47 determines whether the measured temperature T has reached the reuse temperature point Tstd (e). As an example, the reuse temperature point Tstd (e) = 0, 2, 4,..., Floor ((Nstd-1) / 2).
Specifically, the measured temperature T ranges from Tstd (0) ± ΔTstdth, Tstd (2) ± ΔTstdth, Tstd (4) ± ΔTstdth, Tstd ((Nstd−1) / 2) ± ΔTstdth It is determined whether it is in.

  As will be described later with reference to FIG. 8, the range selected in S <b> 16 has no overlapping of temperatures, and thus is used for determination of learning every time the temperature is set to “optimum setting 1”. At this time, it is determined whether or not the measured temperature T has reached the reuse temperature point Tstd (e) (e is an even number).

  It should be noted that, in the determination of reaching the temperature at the reuse temperature point, it is preferable that the matching determination has a width in accordance with the system control cycle and the accuracy of the sensors 20 and 30.

  When the reuse temperature is reached in S18, the reuse temperature determination unit 47 calls the learning result corresponding to the reuse temperature point Tstd (e) from the learning result stored in the storage memory 48 in S15. Then, based on the called learning result (reuse), the measurement value correction unit 45 corrects the output value of the current sensor 20 as the measurement value.

  This control example is mainly applied when the system temperature increase range ΔTsys is wide. Therefore, when the number of corrections Nstd is 2 or more, Nstd / 2 times, and when this is not an integer, (Nstd− 1) Repeat S17 and S18 twice.

  FIG. 7 shows an example of equal allocation when the reuse corresponding to the flow of FIG. 6 is performed.

In FIG. 7, the ambient temperature Tenv at the time of startup = -50 ° C., the system temperature increase range ΔTsys = 130 ° C .− (− 50 ° C.) = 180 ° C. The required learning temperature difference ΔTstdth is set to ± 30 ° C.
From Equation 1, Nstd = floor (ΔTsys / ΔTstdth) = floor (180/60) = floor (6) = 5
In Equation 2, from Nstd = 5, I = 0,1,2,3,4
Tstd (0) = Tenv + ΔTsys / 2-ΔTstdth / 2 × (Nstd-1) + ΔTstdth × I
= -50 + 180 / 2-30 × 1/2 × (5-1) + 30 × 0
= -50 + 90-60 + 0 = -20
In the same way
Tstd (1) =-50 + 90−60 + 30 × 1 = 10
Tstd (2) =-50 + 90−60 + 30 × 2 = 40
Tstd (3) =-50 + 90−60 + 30 × 3 = 70
Tstd (4) =-50 + 90−60 + 30 × 4 = 100
Can be calculated.

  Therefore, when the temperature rises to −20 ° C., 10 ° C., 40 ° C., 70 ° C., and 100 ° C. at the first temperature rise immediately after the start, the learning operation for correction is performed (in FIG. 7, ●, x points) (S13, S14 in FIG. 6 flow).

  On the other hand, if (A) T≈Tstd (e) = 0, 2, 4,..., Floor ((Nstd-1) / 2) is determined as a reuse condition, Tstd (0) = 10 ° C., Tstd (2) = 40.degree. C. and Tstd (4) = 100.degree. C. are the optimum learning temperature points that are the center temperature for sensor characteristic learning during operation (in FIG. 7, ● points) (S13, S15 in FIG. 6 flow). ).

  When the above 10 ° C, 40 ° C, and 100 ° C are set as the optimum learning points, and each time the temperature is reached, learning is performed by calling from the storage memory 48 (reuse), the sensor is re-established centering on the learning temperature point Tstd (e) Since there is no temperature overlap in the range of ± 30 ° C. that is the learning necessary temperature difference ΔTstdth, correction can be performed efficiently with a small number of learning times.

  FIG. 8 is a graph showing the reuse of learning results and the number of learnings in the comparative example and the present invention. Here, a case where 4 × ΔTstdth = ΔTsys is considered. In FIG. 8, ΔTstdth at the time of temperature rise corresponds to the upper half of the arrow.

From Equation 1 above, Nstd = floor (ΔTsys / ΔTstdth) = floor (4/1) = 3
Therefore, the number of corrections excluding the start time is 3, and the learning number including the start time is 4 times.

  In the conventional example shown on the left side of FIG. 8, since the characteristic learning is performed every time when the predetermined temperature passes, the number of times of performing the sensor characteristic learning accompanied by the temporary stop of the power generation of the solar panel (re-learning) increases. As a result, the amount of power generation has decreased.

  On the other hand, the optimum setting 2 shown on the right side in FIG. 8 shows an example in which the same control as in the first embodiment is performed.

  By setting three appropriate learning temperature points Tstd (0), Tstd (1), and Tstd (2) other than the first time (environment temperature at the start), all system operating temperature rise widths Tsys (operating range) ) Can be covered.

  Therefore, even if the temperature fluctuates within the temperature range ΔTsys during system operation, which is the operating range, it is possible to omit setting the operating temperature again by setting an appropriate learning temperature point.

  Therefore, the correction is executed based on the sensor characteristic learning at the time of the first rise, and then the correction is not performed even if the temperature fluctuates. Therefore, the number of learning including the start-up can be suppressed to four times.

  In this way, in the solar cell module system, by setting the optimal learning point for performing sensor characteristic learning of the measurement sensor to the temperature, the number of times of performing sensor characteristic learning with the temporary stop of solar panel power generation is reduced. In addition, a decrease in the amount of power generation can be suppressed.

  Further, the optimum setting 1 in the center of FIG. 8 shows a control example in which the learning value is selected in the second embodiment.

  Furthermore, in the control of the present embodiment, by selecting re-learning for every other stored optimum learning temperature, the entire range can be covered finely without duplication of the learning execution range. Specifically, from the above, assuming that Tstd (0) and Tstd (2), which are an even number of times, are learning points for relearning, the guaranteed range does not overlap when the temperature rises and when the temperature falls. The deviation with high temperature responsiveness of the current sensor can be quickly adjusted with a minimum number of times (half the number of comparisons).

  Therefore, by making it possible to select relearning in this way, it is possible to respond more precisely to the temperature responsiveness of the current sensor, and even if the temperature rise during system operation is wide, the characteristics already learned (deviation) By reusing, the sensor temperature characteristics can be corrected.

  As described above, in this embodiment, in the solar cell module system, by setting the optimal learning point for performing sensor characteristic learning of the measurement sensor to the temperature, the sensor characteristic learning accompanying the temporary stop of the power generation of the solar panel Reduce the number of implementations. Further, by reusing already learned characteristics, it is possible to appropriately correct the sensor temperature characteristics without a temporary stop.

  As mentioned above, although the solar system was demonstrated by embodiment, this invention is not limited to the said embodiment, A various deformation | transformation and improvement are possible within the scope of the present invention.

1,2 Power generation system (solar cell module system)
10 Solar panel 20 Current sensor (voltage sensor, measurement sensor)
30 Temperature sensor 40,400 Measurement value control unit (correction unit)
41 learning condition calculation unit 42 learning temperature storage memory 43 learning temperature determination unit (temperature determination unit)
44 Sensor characteristic learning unit 45 Measurement value correction unit (correction unit)
46 Reuse Temperature Selection Unit 47 Reuse Temperature Determination Unit 48 Storage Memory 50 Power Generation Control Unit (Control Unit)
60 Power conversion unit (control unit)
70 Supplier
Number of learnings excluding Nstd startup
Tenv ambient temperature ΔTstdth Measurement sensor (current sensor) relearning required temperature difference ΔTsys Temperature rise range during solar cell module system operation (temperature rise range during system operation, operating range)
Tstd (i) Optimal learning temperature point that becomes the center temperature for sensor characteristic learning during operation
Tstd (e) Reuse temperature point

Claims (1)

Solar panels with multiple solar cells,
A measurement sensor for measuring an output value of current or voltage output from the solar panel;
A temperature sensor for measuring the temperature of the measurement sensor;
A correction unit that corrects an output value measured by the measurement sensor based on the temperature measured by the temperature sensor;
A sensor characteristic learning unit for performing sensor characteristic learning in order to correct an output value measured by the measurement sensor;
A learning condition calculation unit that calculates a learning temperature point and the number of learning times, which is a temperature at which the sensor characteristic learning is performed;
A temperature determination unit that determines whether or not the temperature of the measurement sensor measured by the temperature sensor has reached the learning temperature point, and a solar cell module system comprising:
In the sensor characteristic learning, by temporarily disconnecting the measurement sensor and the solar panel, a current value or a voltage value generated by a constant current source or a constant voltage source provided in the vicinity of the measurement sensor, and Comparing the generated current value or voltage value with the output value measured by flowing or applied to the measurement sensor to grasp the deviation,
The correction unit corrects an output value measured by the measurement sensor based on the deviation,
When the sensor characteristic learning unit is performing sensor characteristic learning at start-up at the ambient temperature Tenv at start-up, the learning condition calculation unit determines whether the temperature rise range ΔTsys during system operation and the measurement sensor According to the required learning temperature difference ΔTstdth, the learning number Nstd except the start-up time and the optimum learning temperature point Tstd (i) that becomes the center temperature of the sensor characteristic learning during operation;
Calculate the learning count Nstd = (ΔTsys / ΔTstdth) by rounding it down to the nearest lower positive number,
Set the optimal learning temperature point “Tstd (i) = Tenv + ΔTsys / 2−ΔTstdth / 2 × (Nstd−1) + ΔTstdth × i” (i = 0, 1, 2,... (Nstd−1))
When the temperature determination unit determines that the temperature measured by the temperature sensor has reached the optimum learning temperature point Tstd (i), the sensor characteristic learning unit uses the optimum learning temperature point Tstd (i). A solar cell module system that performs sensor characteristic learning during operation.

Images