A new maximum power point tracking controller for photovoltaic power generation

This paper proposes a new maximum power point tracking (MPPT) method in combination with one-cycle control for photovoltaic power generation. This control scheme is based on the automatic current-adjusting feature of one-cycle control. The output current of the inverter can be adjusted according to the voltage of the photovoltaic (PV) array so as to extract the maximum power from it. In the mean time, one-cycle control guarantees that the output current is in the same shape of and in phase with the grid voltage. All these are accomplished in one power stage and a simple control circuit. No detection and calculation of power are needed. Compared with previously proposed approaches, this method is much more efficient and more cost-effective and yet exhibits excellent performance. The principle is explained qualitatively and extensive experiments have been carried out to verify and validate the proposed method.


INTRODUCTION
Photovoltaic (PV) power generation is very desirable since it is renewable and does not contribute to pollution or global climate change. PV is especially attractive for applications in where sunshine is available for most of the year. The commercial viability of PV power generation, however, heavily relies on further improvement of conversion efficiency and reduction of the cost. The effort presented in this paper aims to realize input maximum-power-point tracking (MPPT) and output sinusoidal current injection to the utility grid with a simple and cost effective power electronic solution.
Since the output power of a PV array varies according to the sunlight conditions, atmospheric conditions, including cloud cover, local surface reflectivity, and temperature, MPPT is necessary in order to extract the maximum power from the array at all times. Typically, a PV system employs two-stages of power conversion to transfer the solar power to grid lines: a DC/DC converter is controlled so as to track MPP; and a DC/AC inverter produces a sinusoidal current which is in phase with the grid voltage. This type of system is large, heavy, and expensive, with many components that lead to increased maintenance and reliability concerns. Various approaches have been reported to implement MPPT. The Perturb and Observe (P&O) method needs to calculate dP/dV to determine the maximum power point (MPP) [1,2]. Though it is relatively simple to implement, it can't track the MPP when the irradiance changes rapidly; and it oscillates around the MPP instead of directly tracking it. The Incremental Conductance method can track MPP rapidly but increases the complexity of the algorithm, which employs the calculation of dI/dV [3]. Though this method can get accurate MPPs, a digital signal processor (DSP) or a microprocessor is usually needed for these complex calculations. The Constant Voltage method [4], which uses 76% open circuit voltage as the MPP voltage, and the Short-Circuit Current method [5] are simple, but they do not always accurately track MPPs.
A single-stage PV generation system was presented by Liang et al. [6], but it needs nontrivial calculations for MPPT and the control loop is complex.
In this paper, a new MPPT control method integrated with One-cycle control is presented. There is only one power conversion stage that realizes both MPPT control and dc-ac inversion. With the proposed controller, the new PV power generation system features:

1) Constant switching frequency;
2) Low output current harmonics and PF≈1; 3) Accurate maximum input power point tracking; 4) A simple main circuit with one stage power conversion; 5) A simple controller that only needs some linear components, no DSP, and no multipliers; 6) Potential for low cost and high efficiency.
In the following sections, the basic principle of One-cycle control is introduced. Then the new MPPT control method is explained qualitatively. Finally, the experimental results are presented to prove the concept and validate the proposed method. Fig.1 shows the PV power generation system that is comprised of a PV array, a DC bus buffer capacitor, and an H-bridge inverter. The output is directly connected to the utility grid (grid voltage V o ), and the power flows from the PV array to  For a buck converter:

II. POWER STAGE CIRCUIT AND ONE-CYCLE CONTROL
where, D is the duty ratio.
The control goal is to force the output current i o to follow the grid AC voltage V o . Thus, a unity power factor can be achieved. The control goal equation is: where R e is the emulated resistance, which is related to the output power level. Equation (2) can be rewritten as: where K 1 and K 2 are introduced as positive constants that determine the power level: K 1 limits the maximum output power, and K 1 & K 2 together determine the real output power. Considering the current sensing resistance R s , (3) becomes: , equation (4) becomes: where, K is an introduced constant, and V m (V m >0) is constant during each line cycle. The unity power factor of the system can be achieved by controlling the current to satisfy the control key equation (5), which can be implemented by One-cycle control as shown Fig. 3.  The output power of the inverter P o is a function of two parameters: K determines the up-bound of i o , which limits the maximum output power; and K and V m together determine the real output power. Normally, K is fixed according to the output power level. MPP tracking can be realized by automatically adjusting V m according to environmental factors.
In Fig. 3, during the period between 0 and DT s of each switching cycle, where, Thus, a small change in the PV array voltage V g causes V m to change sufficiently, leading to a change in the output power P o . By this means, the circuit can work stably at the MPP and track it automatically.  If the operating point is first set at the maximum power point (MPP 1 ) and a small perturbation moves the MPP left into Region I, V g decreases. According to (6), V m increases and so i o decreases significantly. Since V o doesn't change, the output power P o of the inverter decreases significantly while P g doesn't drop much for this same perturbation. Since the input power is larger than the output power at the DC bus capacitor C, V g tends to increases, which pulls the operating point back to MPP 1 . On the contrary, if the perturbation makes the operating point move to the right of MPP 1 into Region II, V g increases. This causes V m to decrease, and initiates large increases in i o and P o while P g doesn't increase. As a result, V g on the capacitor C tends to decreases, which pulls the operating point back to MPP 1 again.
I II Therefore, the system design inherently stabilizes the operating point at the MPP.
If the P-V curve moves up (for example, when the solar irradiance rises), P g will increase while P o remains unchanged because V g doesn't change immediately. Then the voltage V g on C will rise, which results in an increase in P g . So the operating point moves up and to the right, towards a new MPP (for example MPP 2 ). On the contrary, if the P-V curve moves down (for example, when the solar irradiance drops), P g will decrease while P o remains unchanged. In this case V g will fall, which results in the decrease of P g . As a result, the operating point moves down and to the left, towards another MPP (for example MPP 3 ). In this way, the MPPT controller can automatically adjust the operating point towards MPPs according to the change of environmental factors. Fig. 6 presents an example of how this MPP tracking scheme works.

IV. EXPERIMENTAL RESULTS
Based on the above theoretical analysis, a 600W experimental system (Fig. 7) was designed and implemented as follows. Fig. 7 presents a photograph of the system that was installed on the rooftop of the Engineering Laboratory Facility at University of California, Irvine. The system is comprised of 8 panels installed at a 45% angle of incidence and facing roughly 15 o east of due south. This configuration was decided based upon building orientation, yearly average angle of direct solar incidence, and needs for sturdy mounting of the support structure. This orientation is not optimized for the months of October and November in Southern California, when the experiments were conducted.
Specifications of the PV panels used in the experiments are presented in Table I. Note that each of the PV array modules is specified to produce 75W of peak power with direct irradiance levels of 1000W/m 2 of spectrum AM 1.5 when the cells are at 25 o C.  Fig. 9 shows a representative sample of the output voltage V o and current i o (tested at 9:25AM on the same day) of the inverter system described above. Because the typical output of the PV panels is not sufficient to supply a voltage that can be directly inversed to the line voltage by the H-bridge inverter (the range of voltage experienced during the experimental period was 0-165V), the output was connected to the AC power grid via an autotransformer in this experiment. The resistance of the autotransformer causes some distortion to the waveform presented in Fig. 9. In Fig. 9

V. CONCLUSIONS
In this paper, a new MPPT control method combined with One-cycle control is proposed. The MPPT controller can automatically adjust the operating point of the PV system to the MPP according to changes in environmental factors. In the mean time, the output current of the inverter is sinusoidal with unity power factor. The system is simple and efficient comprised of only one power stage and one simple controller to implement two functions: MPPT and dc/ac inversion. Due to this simple control circuit, this method shows promise for reducing the cost and complexity and increasing the efficiency for commercial PV systems. The approach could also be applied to fuel cells or other distributed generation technologies that also need simple and efficient inverter technology. Output voltage Vo (curve 1, 100V/div) and io (curve 2, 10A/div)