Fault Detection Algorithms for Achieving Service Continuity in Photovoltaic Farms A simulated 250-kW PV power plant was utilized to create training and testing datasets of PV fault cases. The PV farm and its simulation are further discussed in Appendix A. Three fault types and normal operation (free-of-fault state) are defined. The default sets, as shown in Appendix A, are as follows. From the figure presented in the Appendix section, the fault cases F1, F2, and F3 describe a string fault (tested on string 1), string-to-ground fault (tested on string 1), and string-to-string fault (tested between strings 1 and 2), respectively. Training and testing datasets were built. The training dataset included 600 instances, each with 30 features and one column for classes or categories. Tab. 1 shows that the dataset included 100 (16.67%) free-of-fault cases, 153 cases (25.5%) of string faults, 149 cases (24.83%) of string-to-ground faults, and 198 cases (33%) of string-to-string faults. The total simulation time was 0.4 s, and a fault was assumed to occur at 0.2 s. In the training dataset, all measurements were taken after the fault occurred in the period from 0.2 s to 0.4 s. The testing dataset contained 50 instances. Measurements were taken in the period from 0.1 s to 0.3 s, with transient time from 0.1 s to 0.2 s, and faults occurring from 0.2 s to 0.3 s. for more information about this dataset see the following link: https://www.techscience.com/iasc/v30n2/44023 Please Cite This Article S. S. M. Ghoneim, A. E. Rashed and N. I. Elkalashy, "Fault detection algorithms for achieving service continuity in photovoltaic farms," Intelligent Automation & Soft Computing, vol. 30, no.2, pp. 467–479, 2021. video tutorial (see project 1) https://youtu.be/wpgQY5f2hOo presentation (see project 1) https://www.slideshare.net/AmrRashed3/machine-learning-workshop-using-orange-datamining-framework codes : https://github.com/amrrashed/Fault-Detection-Dataset-in-Photovoltaic-Farms