We know that when the sunshine intensity and ambient temperature change, the output voltage and current of photovoltaic cells change in a nonlinear relationship, and its output power also changes. Moreover, when photovoltaic cells are applied to different loads, the output power of the photovoltaic system is reduced due to the mismatch between the output impedance of photovoltaic cells and the load impedance. At present, an effective way to solve this problem is to add a switch conversion circuit between the output end of the photovoltaic cell and the load, and use the principle of impedance transformation of the switch conversion circuit to make the equivalent impedance of the load follow the output impedance of the photovoltaic cell, so that the photovoltaic cell output power is maximum. Common maximum power tracking methods are as follows: ① power-matching scheme; ② curve-fitting technique; ③ Perturb-and-observe method; ④ Incremental conductance algorithm.

Method ① needs to obtain the output characteristics of the solar array, and can only be applied under specific radiation and load conditions, so there are certain limitations. Method ② requires that the characteristics of the solar energy battery array be measured in advance and described by detailed mathematical functions. However, this method fails for changes in characteristics due to lifetime, temperature, and damage of individual cells. Method③ is a selection process, without knowing the characteristics of the solar array, so it is a widely used method; The disadvantage is that the operating point of the system cannot be stabilized at the maximum power point due to the intervention of disturbance. The method ④ solves this problem. The calculation is accurate and can prevent the misjudgment of the working point.

According to the maximum power criterion, the tracking methods can be as follows.

(1) The output voltage invariant method holds that the output voltage of photovoltaic cells changes very little when conditions such as illumination change, so it only needs to find a certain operating point to maintain its output voltage near the maximum power point. Obviously, the error of this method is relatively large.

(2) power calculation method of photovoltaic cells through testing the output voltage and current, so as to calculate the output power at this time, the approximation by comparison with the previous value of the maximum power point, sometimes join in this method of artificial neural network and other intelligent control, adaptive control and fuzzy control algorithm, used to judge the power change trend. The neural network is used to determine the output voltage of the DC/DC BOOST converter to obtain the maximum power output of the photovoltaic cell. Once the neural networks are trained, they can be used in dedicated systems. Predictive current control can be achieved using fuzzy logic, that is, to calculate the required on-cycle of each inverter bridge arm, which drives the corresponding line current to the reference value in one switching cycle. This kind of method has the advantage of direct control mode, but it also has the problem of complex calculation.

(3) Current optimization method This method is mostly used in the DC power generation system of battery charging. Because the battery voltage is relatively stable, the maximum power output can be obtained only by ensuring the maximum output current of the photovoltaic cell. The advantage of this method is less detection, but there is still the problem of output disturbance, and it is not suitable to use when the output voltage fluctuation is large.

(4) Ripple disturbance method is also known as parasitic capacitance method, that is, the ripple generated by switching action is used to replace the disturbance artificially injected. It is easy to measure and handle, but there are also some adverse factors: The working performance depends on the SNR. If a large filter is used in the line, this method will fail, and a high switching frequency means a small ripple, which is unfavorable for the judgment of the working state. The judgment of the working state is affected by the dynamic characteristics of the photovoltaic cell and also by the capacitance.

From the point of view of the maximum power tracking circuit, the following circuits are mainly adopted.

(1) BUCK circuit Although BUCK circuit is more efficient in common applications, it is rarely used in maximum power tracking circuits. One reason is that most converters are required to operate in continuous current mode and output as much power as possible.

(2) BOOST circuit This kind of circuit is widely used.

(3) SEPIC and CUK circuits, when operating in discontinuous inductor current mode or capacitor voltage mode, have comparable or reverse comparable resistance characteristics, which just meet the needs of maximum power output regulation. The maximum power point can be located by comparing the average of the AC component and the voltage at the photovoltaic cell end by modulating a small signal sinusoidal distortion to the switching signal.

(4) The inverter type is characterized by the integration of the maximum power tracking in the inverter control, without additional conversion circuit, limited to the photovoltaic power generation inverter system.

There are also references to improve the DC/DC converter circuit, mainly to add the feedforward correction link, so that the converter works in current mode. This method has the following advantages: (1) the battery is generally used as the load, and only the output inductor current is controlled; ② Multiple converters can be used in parallel. In addition, there are literatures to realize the maximum power tracking by hardware circuit, for example, using automatic oscillator to adjust some parameters of DC/DC converter (such as the conduction period), so that the photovoltaic cell output work at the maximum power point, the commonly used methods are voltage feedback and power feedback.