We propose a real-time information propagation arithmetic neural network (PANN) that minimizes the loss of power generation output of the system in the event of sudden changes in the module due to strong external typhoons or earthquakes at the solar power generation facility site. In addition, we propose a new double-sided module reflector that can reduce the local loss of power generation efficiency of the single-sided module reflector that is currently widely distributed, as well as the environmental pollution and inconvenience of maintenance work of the existing double-sided module. We present a computational network that can detect the faulty solar panel in real-time by checking the fault status of the installed solar panel and using a real-time computation method through a node-to-node diffusion method. In particular, this method recognizes the power loss part due to sudden changes in the module in real time and can take emergency measures for various nonlinear field facilities through a neural structure that finds the optimal distance up, down, left, and right. To confirm the characteristics of the loss reduction control of the field facility, we confirmed that the system was configured as a 7-degree-of-freedom control model using the PANN neural network learning structure method and improved the power generation output. PANN (Propagation Arithmetic Neural Networks) and various module systems are proposed for the real-time recovery of faulty solar panels and improving module system efficiency.
