Press "Enter" to skip to content

Artificial neural networks and fuzzy logic for fault detection in PV systems

Researchers in China have developed a fault detection method for PV systems that combines both neural networks and fuzzy logic principles within a single framework. It considers seven input variables and an output variable.

Researchers from the Zhengzhou University in China have created a new electrical fault detection system for PV systems by using the Adaptive Neuro-Fuzzy Inference System (ANFIS) methodology, which is an artificial neural network (ANN) technology based on the so-called Takagi–Sugeno fuzzy inference system.

The latter is a method to map an input to an output using fuzzy logic. Through this process, values of the input vector can be interpreted and, on the basis of some sets of fuzzy rules, corresponding values to the output vector can be assigned. The ANFIS method combines both neural networks and fuzzy logic principles within a single framework and, according to the Chinese group, is more accurate than fuzzy logic and artificial neural networks operating separately.

“Combining the ANN and fuzzy-set theory can provide advantages and overcome the disadvantages in both techniques,” the academics stated. “The ANFIS model can be trained without relying solely on expert knowledge sufficient for a fuzzy logic model.”

The model comprises seven inputs and one output. The seven input variables are I1I1″> (Amperes), V1″> V1 (Volts), I2″>I2(Amperes), V2″> V2 (Volts), Irradiance (Klux), Temperature (°C), and Weather (sunny/cloudy), and the output variable is the fault state of the PV system.

The proposed ANFIS approach was tested on a 1.8 kW PV system’s experimental setup with both grid partitioning (GD) for hourly solar radiation forecasting and subtractive clustering (SC), which is an algorithm used to generate the tuned membership functions automatically in accordance to the domain knowledge and is the base to set up adaptive network inference systems.

Popular content

The methodology was found to be able to accurately detect PV fault states. “The proposed ANFIS SC methodology is superior to the ANFIS GP technique in accurately detecting PV fault states,” the researchers concluded, adding that the new technique was also able to accurately track the experimental data in comparison to soft-computing techniques.

The detection system is presented in the paper A smart fault detection approach for PV modules using Adaptive Neuro-Fuzzy Inference framework, published in Energy Reports.

This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: [email protected]

Source: pv magazine