System Of PV Power Station Data Collection And Monitoring Based On Diagonal Recurrent Neural Network

Solar energy is a huge amount of renewable energy, solar energy reaching the Earth surface every day is equivalent to hundreds of millions of barrels of oil burning energy. Developing and using of solar energy has great strategic significance, it is very important from economic and social path of sustainable development path and the protection of human life on Earth to look at the height of the ecological environment, or from the special purpose to solve practical problems of energy supply. As the photovoltaic power generation has many advantages unmatched by traditional energy sources, it has been recognized one of the best ways of energy supply as human beings, but because of the high cost of photovoltaic power generation system equipment, the main equipment such as solar cells and batteries of damage or reduced life will directly reduce the efficiency of photovoltaic power generation, so the photovoltaic power generation system of data acquisition and monitoring is inevitable.In this paper, AT89C51 microcontroller is the core of data collection and testing devices for photovoltaic power generation system. After collecting and monitoring the solar cell, battery voltage and current, inverter output voltage and current, ambient temperature, the battery plate surface temperature, the control room's temperature of photovoltaic power generation system, through the LCD screen displays the test results and ultimately forms voltage, current, temperature characteristic curve equivalent in a period of time.Since these curves are non-linear curve, so this paper introduces the diagonal recurrent neural network for modeling and simulation. This paper introduces the concept, development, research content and application of neural networks, nonlinear systems have a more detailed introduction,then analyzes neural network system identification, leads to elaborate on the diagonal recurrent neural network in this paper. Because of the diagonal recurrent neural network hidden layer nodes with self-feedback function, so the model has higher accuracy. DRNN of this article is trained by BP learning algorithm, although this algorithm has slow convergence of the defect, but their algorithm is simple and easy to train, so it is widely used. Simulation results show that DRNN-based photovoltaic power plant data acquisition monitoring system is relatively successful.