TWI481989B - A system of solar power generator with power tracker - Google Patents

Description

Solar power system with power tracker

The present invention relates to a solar power generation system, and more particularly to a solar power generation system having a power tracker.

Under the general consideration of the energy crisis in the world, all parties are not committed to finding alternative energy sources to maintain the supply of people's livelihood and continue the development of industry. Among them, the use of solar energy as an alternative energy source for power generation, in addition to eliminating the exhaustion of the use of power sources, can reduce the emissions generated by power generation operations, and help maintain the natural environment.

For solar power generation systems, different output voltages, output currents and temperatures will be used in different sunshine environments, and different output voltages, output currents and temperatures also have different output power curves and higher outputs. Power, which means that it has a high photoelectric conversion efficiency. If the output power of the solar power generation system can be continuously tracked and the operation of the control circuit is maintained, the solar power generation system can maintain a higher output power state, and it can receive better power generation performance.

In the maximum power tracking method of the conventional solar power generation system, it is common to use the disturbance observation method for tracking. The disturbance observation method periodically changes the load of the solar power generation system, and observes the power change caused by the load variation, and then determines the fluctuation direction of the load of the next cycle by the power change, so that the solar power generation system can Maintain large output power under different power generation conditions. However, by observing the observation method to observe the power variation of the solar power generation system, it is necessary to perform the output work by constantly changing the load. Rate detection, even if the solar power system has been operating at the maximum power point, the load change will not stop, and when the load oscillates back and forth around the maximum power point, it will cause additional power loss, thereby reducing the solar energy. The output power of the power generation system. In view of this, a better power tracker must be provided for the solar power generation system.

It is a primary object of the present invention to provide a solar power generation system having a power tracker that maintains the solar power generation system at a higher output power state without additional power loss.

In order to achieve the foregoing object, a solar power generation system with a power tracker of the present invention includes: a solar panel for outputting direct current power, the solar panel having a temperature signal having a voltage signal and a current a signal; a DC/DC converter coupled to the solar panel to receive the DC power, and converting the DC power to output an AC power having an AC voltage signal; and a power tracker coupled The solar panel and the DC/DC converter, the power tracker comprises: a type of neural network system, receiving the voltage signal, the current signal, the temperature signal and the DC conversion voltage signal, and performing a neural operation to output a reference The voltage signal, wherein the neural network system has an input layer, a type layer, a sum layer and an output layer, and the input layer receives an input vector formed by the voltage signal, the current signal and the temperature signal. And assigning the input vector to each operation node of the type layer; the type layer has several operation nodes Each of the computing node are present having the same vector, after each of the computing node receiving the input vector, the input vector is calculated based with From the square of one of the sample vectors, the square is combined into an excitation function to obtain a plurality of type output values; the sum layer includes a first sum node and a second sum node, the first sum node has a weight value, after the first sum node receives the plurality of type output values, performing weighted summation operations on the plurality of type output values and the weight value to obtain a first sum value, the second sum node After receiving the plurality of type output values, the plurality of type output values may be summed to obtain a second sum value; the output layer receives the first sum value and the second sum value, and The first sum value is divided by the second sum value to obtain the reference voltage signal; a controller receives the reference voltage signal and the DC converted voltage signal, and compares the reference voltage signal and the DC converted voltage signal according to the reference voltage signal As a result, a control signal is output to the DC/DC converter.

The solar power generation system of the present invention has a power tracker, wherein the neural network system further has a learning module, and the learning module receives the reference voltage signal and the DC conversion voltage signal, and calculates the reference voltage signal and One of the DC-converted voltage signals has a mean square error, and the mean square error is used as one of the excitation parameters to smooth the parameters.

The solar power generation system with power tracker of the present invention, wherein the excitation function is as follows:

Where σ is the smoothing parameter and D is the square of the input vector and the sample vector of each computing node.

The solar power generation system with a power tracker of the present invention, wherein the DC/DC converter has a receiving end and an output end, the receiving end and the input end The output terminal includes a high level line, a low level line, an input voltage stabilizing capacitor, a power switch module, an output voltage stabilizing capacitor, an inductor and a power diode, and the input voltage stabilizing capacitor, The power switch module and the output voltage stabilizing capacitor are connected in parallel between the high level bit line and the low level bit line, the inductor is disposed on the high level bit line, and the two ends of the inductor are respectively connected to the input voltage stabilizing capacitor and power Switch module. The power diode is disposed on the high level line, and the anode of the power diode is connected to the power switch module, and the cathode of the power diode is connected to the output voltage stabilizing capacitor.

The solar power generation system of the present invention has a power tracker, wherein the DC/DC converter is coupled to a DC/AC converter, and the DC/AC converter is coupled to a filter circuit, and the filter circuit is coupled to an AC load.

The above and other objects, features and advantages of the present invention will become more <RTIgt;

"Coupling" as used throughout the present invention refers to a connection between two elements through direct or indirect connection, such that signals or currents are transmitted between the two elements.

Referring to FIG. 1 , the solar power generation system with power tracker of the present invention comprises: a solar panel 1 , a DC/DC converter 2 and a power tracker 3 . The DC/DC converter 2 is coupled to the solar panel 1 , and the power tracker 3 is coupled to the solar panel 1 and the DC/DC converter 2 .

The solar panel 1 can be photoelectrically converted after being received by light to output Current power, wherein the DC power has a voltage signal and a current signal. The solar panel 1 preferably has a temperature sensor for sensing the temperature of the solar panel and outputting a temperature signal.

Referring to FIGS. 1 and 2, the DC/DC converter 2 has a receiving end 21 and an output end 22, and the receiving end 21 receives the DC power and performs boost conversion on the unstable DC power. And outputting a stable DC conversion power from the output terminal 22. Wherein, the DC conversion power has a DC conversion voltage signal.

In this embodiment, the receiving end 21 and the output end 22 of the DC/DC converter 2 include a high level line 23, a low level line 24, an input voltage stabilizing capacitor 25, and a power switch module. 26, an output voltage stabilizing capacitor 27, an inductor 28 and a power diode 29, the input stabilizing capacitor 25, the power switch module 26 and the output stabilizing capacitor 27 are connected in parallel to the high level line 23 and the low level Between bit lines 24. The inductor 28 is disposed on the high level line 23, and the two ends of the inductor 28 are respectively connected to the input voltage stabilizing capacitor 25 and the power switch module 26. The power diode 29 is disposed on the high-level line 23, and the anode of the power diode 29 is connected to the power switch module 26. The cathode of the power diode 29 is connected to the output voltage regulator 27.

The power tracker 3 has a type of neural network system 31 and a controller 32. The neural network system 31 receives the voltage signal, the current signal, the temperature signal and the DC conversion voltage signal of the solar cell, and performs a neural-like operation to output a reference voltage signal.

Referring to FIG. 3, the neural network system 31 has an input layer 311, a type layer 312, a sum layer 313, an output layer 314, and a learning module 315.

The input layer 311 receives an input signal formed by the voltage signal, the current signal and the temperature signal, and distributes the input vector to each of the operation nodes of the type layer 312.

The type layer 312 has a plurality of operation nodes, each of the operation nodes has the same local vector. After receiving the input vector, each of the operation nodes calculates the square of the input vector and the respective sample vector, and then squares the square. Combine the excitation function to obtain several types of output values.

The excitation function of this type of layer 312 is as follows:

Where σ is a smoothing parameter, which is a constant greater than 0, and D represents the square of the input vector and the sample vector of each computing node.

The calculation of this type of layer 312 is as follows:

among them Represents the input vector, A sample vector representing the i- th compute node.

The summation layer 313 includes a first sum node and a second sum node. The first sum node has a weight value. After receiving the plurality of type output values, the first sum node can output the plurality of patterns. The value is weighted and summed with the weight value to obtain a first sum value, and after the second sum node receives the plurality of type output values, the plurality of type output values may be summed to obtain a The second sum value.

The equation of operation for the first sum node is as follows:

Where S ₁ represents the first sum value and Y represents the weight value.

The operation equation of the second sum node is as follows:

Where S ₂ represents the second sum value.

The output layer 314 receives the first sum value and the second sum value, and divides the first sum value by the second sum value to obtain the reference voltage signal.

The operational equation of the output layer 314 is as follows:

among them Represents the reference voltage signal.

The learning module 315 receives the reference voltage signal and the DC conversion voltage signal, and calculates a Mean Square of Error (MSE) of the reference voltage signal and the DC conversion voltage signal, and uses the mean square error As a smoothing parameter of the excitation function.

The calculation formula of the learning module 315 is as follows:

Where Q represents the number of samples, V _MPPT represents the reference voltage signal, and V _dc represents the DC conversion voltage signal.

Further, by setting the learning module 315, the reference voltage signal and the direct current can be calculated according to the actually outputted DC converted power. Converting the mean square error of the voltage signal, and using the minimum mean square error as the smoothing parameter of the excitation function, and continuously updating the reference voltage signal, thereby achieving better control effect.

The controller 32 receives the reference voltage signal and the DC conversion voltage signal, and outputs a control signal to the DC/DC converter 2 according to the comparison result of the reference voltage signal and the DC conversion voltage signal.

In this embodiment, the controller 32 can be a circuit module such as a comparator, and compare the reference voltage signal and the DC conversion voltage signal to output the control signal to the power switching mode of the DC/DC converter 2. Group 26, thereby controlling the switching period of the power switch module 26 to operate at a maximum output power point under different operating conditions, thereby causing the DC/DC converter 2 to output DC converted power at a maximum power point.

Referring to FIG. 1 again, in order to make the power output of the present invention applicable to loads of different power requirements, the DC/DC converter 2 can be connected to the DC/AC converter 4 to convert the DC conversion power. For an AC power. The DC/AC converter 4 is coupled to a filter circuit 5, which is coupled to an AC load 6. The DC conversion power is converted to AC power and then filtered by the filter circuit 5 for filtering. This AC power can be smoothly used for the AC load 6.

The neural network system 31 such as the power tracker 3 can perform power tracking according to the actual output DC conversion power, and adjust the DC conversion power outputted by the next timing, thereby enabling the solar power generation system to operate. Maximum power point. Moreover, the present invention does not perform power tracking by means of load oscillation, so that there is no power loss due to load oscillation, and the solar power generation system can be maintained at a high output power state, Has the power of additional power loss.

While the invention has been described in connection with the preferred embodiments described above, it is not intended to limit the scope of the invention. The technical scope of the invention is protected, and therefore the scope of the invention is defined by the scope of the appended claims.

〔this invention〕

1‧‧‧ solar panels

2‧‧‧DC/DC Converter

21‧‧‧ Receiver

22‧‧‧ Output

23‧‧‧High standard line

24‧‧‧low standard line

25‧‧‧Input Stabilizing Capacitor

26‧‧‧Power switch module

27‧‧‧ Output Stabilizing Capacitor

28‧‧‧Inductance

29‧‧‧Power diode

3‧‧‧Power Tracker

31‧‧‧ class neural network system

311‧‧‧Input layer

312‧‧‧Type layer

313‧‧‧Total layer

314‧‧‧ Output layer

315‧‧‧ learning module

32‧‧‧ Controller

4‧‧‧DC/AC converter

5‧‧‧Filter circuit

6‧‧‧AC load

Figure 1 is a block diagram of a solar power generation system with a power tracker of the present invention.

Fig. 2 is a circuit diagram of a DC/DC converter circuit of a solar power generation system with a power tracker according to the present invention.

Figure 3: Schematic diagram of a neural network system such as a solar power generation system with a power tracker of the present invention.

1‧‧‧ solar panels

2‧‧‧DC/DC Converter

3‧‧‧Power Tracker

31‧‧‧ class neural network system

32‧‧‧ Controller

4‧‧‧DC/AC converter

5‧‧‧Filter circuit

6‧‧‧AC load

Claims (5)

A solar power generation system with a power tracker includes: a solar panel for outputting direct current power, the solar panel having a temperature signal having a voltage signal and a current signal; a DC/DC converter The solar panel is coupled to receive the DC power, and the DC power is boost-converted to output a DC-converted power having a DC-converted voltage signal; and a power tracker coupled to the solar panel And the DC/DC converter, the power tracker comprises: a type of neural network system, receiving the voltage signal, the current signal, the temperature signal and the DC conversion voltage signal of the solar cell, and performing a neural operation to output a reference a voltage signal, wherein the neural network system has an input layer, a type layer, a sum layer and an output layer, and the input layer receives an input vector formed by the voltage signal, the current signal and the temperature signal of the solar cell. And assigning the input vector to each of the operation nodes of the type layer; the type layer has Each of the operation nodes has the same local vector. After receiving the input vector, each of the operation nodes calculates the square of the input vector and the respective sample vectors, and then substitutes the square into an excitation function to Obtaining a plurality of type output values; the summation layer includes a first sum node and a second sum node, the first sum node has a weight value, and the first sum node receives the plurality of type output values, Performing a weighted sum operation on the plurality of type output values and the weight value to obtain a first sum value, and after receiving the plurality of type output values, the second sum node may input the plurality of patterns The value is summed to obtain a second sum value; the output layer receives the first sum value and the second sum value, and divides the first sum value by the second sum value to obtain the reference a voltage signal; a controller receiving the reference voltage signal and the DC conversion voltage signal, and outputting a control signal to the DC/DC converter according to the comparison result of the reference voltage signal and the DC conversion voltage signal. The solar power generation system with a power tracker according to claim 1, wherein the neural network system further has a learning module, and the learning module receives the reference voltage signal and the DC conversion voltage signal. And calculating a mean square error of the reference voltage signal and the DC conversion voltage signal, and using the mean square error as one of the excitation parameters to smooth the parameter. The solar power generation system with a power tracker according to claim 1, wherein the excitation function is as follows: Where σ is the smoothing parameter and D is the square of the input vector and the sample vector of each computing node. The solar power generation system with a power tracker according to claim 1, wherein the DC/DC converter has a receiving end and an output end, and the receiving end and the output end comprise a high level line, a low level line, an input voltage stabilizing capacitor, a power switch module, an output voltage stabilizing capacitor, an inductor and a power diode, the input voltage stabilizing capacitor, the power switch module and the output voltage stabilizing capacitor are connected in parallel Between the high level bit line and the low level bit line, the inductor is disposed on the high level bit line, and the inductor The second end is connected to the input voltage stabilizing capacitor and the power switch module. The power diode is disposed on the high level line, and the anode of the power diode is connected to the power switch module, and the cathode of the power diode is connected to the output voltage stabilizing capacitor. The solar power generation system with a power tracker according to claim 1, wherein the DC/DC converter is coupled to a DC/AC converter, and the DC/AC converter is coupled to a filter circuit, the filter circuit An AC load is coupled.