Solar irradiance estimation and optimum power region localization in PV energy systems under partial shaded condition

Abstract

The efficient operation of PV systems relies heavily on maximum power point tracking (MPPT). Additionally, such systems demonstrate complex behavior under partial shading conditions (PSC), with the presence of multiple maximum power points (MPP). Among the existing MPPT algorithms, the conventional perturb and observe, and incremental conductance stand out for their high simplicity. However, they are specialized in single MPP problems. Thus, due to the existence of multiple MPPs under PSC, they fail to track the global MPP. Compared with the conventional schemes, the modified conventional algorithms, and several existing MPPT variants introduce a trade-off between complexity and performance. To enhance the simplicity of the PV system, it is crucial to adapt the operation of the simple conventional algorithm to scenarios under PSC. To achieve such an adaptation, the power-voltage curve that conventionally admits multiple MPPs under PSC must be converted to an equivalent curve having only a single MPP. To address such a requirement, this paper introduces a novel approach to the fast determination of the MPP. A consistent methodology for reducing the complex multiple MPP problem of PV systems under PSC, to a single MPP objective, is put forward. Thus such reduction enhances the tracking environment for simple conventional MPPT algorithms under partial shading. Studies of the PV array behavior for 735 partial shading patterns revealed an interesting possibility of reducing the classical PV curve to 8.2620% of its actual area. The newly established area is an optimum power region that accommodates a single MPP. To arrive at such a reduction, an intelligent neural network-based predictor, incorporating a cost-effective and reliable solar irradiance estimator is put forward. Unlike existing methods, the approach is free from the direct and expensive measurement of solar irradiance. The predictor relies on the PV array current and voltage only to precisely determine the optimum power region of the PV system.

Keywords: Irradiance estimator; MPPT; Optimal power region; PV; Partial shading conditions.

Graphical abstract

Fig. 1

Single diode model of the PV.

Fig. 2

Operation of the PV array (a) under uniform irradiance condition (b) under partial shading condition.

Fig. 3

Characteristics of the PV array under UIC and PSC (Descriptive example).

Fig. 4

Solar irradiance Estimator under PSC.

Fig. 5

735 different patterns of PSC for three PV modules.

Fig. 6

Distribution of the GMPP for 735 PSC patterns.

Fig. 7

Convergence of the mean square error during the training process.

Fig. 8

Synoptic structure of the proposed system.

Fig. 9

Implementation of the proposed system in Simulink/MATLAB.

Fig. 10

Dynamic response of the estimator for different values of the estimator design parameter (k).

Fig. 11

Response of the estimator under PSC- B

Fig. 12

Localization of the optimal power region under PSC-B

Fig. 13

Response of the estimator under PSC- C

Fig. 14

Localization of the optimal power region under PSC-C

Fig. 15

Response of the estimator under PSC- D.

Fig. 16

Localization of the optimal power region under PSC-D.

Fig. 17

Response of the Estimator under changing PSC patterns.

Fig. 18

Dynamic response of the lower and upper voltage bound for the optimal power regions.

Fig. 19

Distribution of the new area of the P–V curve evaluated from the PV array using 700 PSC patterns.

Fig. 20

Conversion of the multiple power point P–V curve to a single maximum power point P–V curve under PSC.