Investigation On The Ultraviolet Absorption Spectrum Of The Nitrate Solution And The Phosphate Solution

With the social development and the industrial progress, the water pollution gets more seriously. It has become the first environmental governance problem of the world. Not only does the water pollution affect people's production and life seriously, but also it threatens people's health. Many countries in the world have put the governance of water pollution as principal mission. Scholars and researchers have put into how to govern the water pollution. They analyze the causes of pollution, and find many solutions.If the content of the nitrogen and phosphate in the water exceeds certain standard, it will cause serious water pollution. In this paper, the object of research are nitrogen and phosphate. This paper uses the ultraviolet spectrophotometry to analyze. 220nm is the wavelength point of the best linearity. The absorbency is directly proportional to the density of nitrate solution from 0mg/L to 4mg/L through the experiment. After 4mg/L the working curve presents the non-linear trend. Using the monadic linear regression we get the regression line. Its correlation coefficient is 0.9988. At the same time we obtain small standard error of the regression coefficient.Sanitary waste contains many of phosphates. If the sanitary waste is discharged at random, it cause the content of phosphate will be very high in rivers, lakes and the sea. Then it will cause water eutrophication. Because of the instability of phosphate, we convert the experimental data. Then we get the non-linear relationship between absorbance and the concentration of solution. The BP neural network is used to deal with the non-linear problem. We use Matlab language to realize the BP network algorithm. In experiments, the network convergence rate of the Levenberg-Marquardt method is the fastest. After 54 times cycle the method reaches the accuracy. The rate of the quantized conjugate gradient method is the slowest. It doesn't reach the accuracy after 300 times cycle. Using the data fitting we receive 12 times fitting curve. Comparing with BP network algorithm, the data fitting is a complex method. And its result is poor.This paper enriches the application of the UV absorption spectrum detection. And the paper combines with the linear regression method and the BP neural network algorithm. It makes the technology which is about the governing water pollution of UV detection improve and enhance. To further study, the paper provides basis which is about the various causes of water pollution. And it provides scientific reference which is about the technology of photoelectric detection.