Sensores de baixo custo para monitoramento continuo da umidade do solo e radiação solar

Abstract

The implementation of low-cost sensors allows the realization of monitoring studies through spatial temporal distribution of real time data, enhancing a greater number of researches to be carried out. The construction and use of these sensors also have a didactic function, since it allows the student to get more directly involved in understanding the physical characteristics that define the process being studied. In the present dissertation, two types of low-cost sensors are characterized, a solar radiation sensor and a soil moisture sensor, built and improved with the purpose of being used in multiple research and academic practices. The working principle of the sensor for humidity estimation is capacitive-resistive. This sensor reacts to the variation in the electromagnetic properties of the soil, contained between its metallic rods. These properties vary proportionally with changes in the water content present in the soil pores. It was developed a mathematical model that estimates the output frequency of the sensor oscillator circuit based on the probe-soil interaction in order to estimate the percentage of moisture present in the soil. By applying the model in a computer algorithm, it was possible to estimate the volumetric soil moisture with an uncertainty of less than ± 3% at a 95% confidence interval. The algorithm can be used for different soil textures by entering the porosity and the electrical conductivity of the saturation extract of the soil as parameters. With regard to the pyranometer, the measurement of solar radiation is of great importance for estimating the evapotranspiration demand of the atmosphere. Despite its importance, solar radiation has not been measured at many weather stations due to the high cost of this sensor. The pyranometer is based on a silicon photodiode, which induces an electric current proportional to the radiation reaching it. Sixteen sensors were built to evaluate their accuracy in estimating solar radiation and the effect of uncertainties in estimating reference evapotranspiration. The sensors were installed in the open field and in the greenhouse, performing continuous measurements for one year in conjunction with a secondary standard thermopile sensor. During this time the sensors showed no physical deterioration nor changes in standard deviation (0,04 MJ.m-2day-1). The uncertainty in the estimation of ETo by the Penman-Monteith algorithm using the low cost sensor was ± 0,22 mm.day-1 for 95%.