Research On Solar Energy Harvesting And Prediction For Wireless Sensor Network

With the rapid development of electronic technology,the Internet of things,wireless sensor networks and other applications are gradually improving people’s lives.Many scenarios can be monitored by wireless sensor nodes,such low-power devices are usually deployed in large-scale unattended areas.Nodes usually use disposable battery power supply,when the power is exhausted,it is extremely difficult to replace the battery for thousands of nodes,and the continuous work of the whole sensor network will be affected.Harvesting solar energy for nodes is one of the effective ways to solve the problem of frequent battery replacement.At the same time,it is also necessary to predict the solar energy that can be harvested,so that the node can correctly estimate the state of charge of the system combined with the battery power and the energy that can be harvested in the future,so as to provide guarantee for the reasonable work scheduling of the node.Firstly,this paper summarizes the research status of solar energy harvesting and prediction of self-powered wireless sensor nodes at home and abroad,and analyzes the components of the power structure of self-powered wireless sensor nodes.On this basis,the following research contents are completed:(1)For the lack of cost,performance,complexity and other aspects of comparison between the solar energy harvesting circuits of wireless sensor network.Through the actual construction of the existing solar energy harvesting circuit,the test is carried out under outdoor natural light conditions,and according to the test results,the energy harvesting efficiency,complexity,power consumption,cost and applicability of each circuit are analyzed.(2)For the problem of complex structure and high cost of existing harvesting circuit.The hardware cost and complexity are improved.A suboptimal minimalist solar energy harvesting circuit structure is proposed,and the feasibility analysis and actual test of the proposed circuit structure are carried out.It is verified that the circuit can still obtain considerable energy harvesting efficiency at low cost.(3)An energy harvesting circuit structure for solar panel array is proposed.By arranging a plurality of solar panels in the shell of the node,a solar panel array is formed to supply power for the node.This method can make full use of the outer surface of the node and provide more energy for the node with higher energy consumption.By analyzing the theoretical output characteristics of solar panel array,the corresponding hardware circuit structure and software algorithm are designed to solve the problem of low energy harvesting efficiency.(4)For the problem that the existing energy prediction algorithms have large errors when the weather changes alternately.A prediction method of solar energy based on Kmeans clustering is proposed.Using historical energy data to build a classification database,in order to distinguish the energy that can be obtained under different weather conditions.By selecting similar weather samples,we can accurately predict the change of energy when the weather changes alternately.On this basis,the influence of the number of similar weather samples on the prediction error is discussed.