Research On Water Pollution Early Warning Method Based On Cloud Platform

Although the traditional chemical method has high water quality detection accuracy,the detection cycle is long,requires professional operations,and the chemical reagents used are prone to secondary pollution.In recent years,UV-visible spectroscopy has been widely used in water quality detection.It has the advantages of fast detection speed,no secondary pollution,fingerprint detection,and traceability of pollution sources and so on.And it has become a research focus in the field of water quality detection.In the UV-visible spectroscopy water quality detection system,in order to improve the accuracy and stability of the entire detection system,higher requirements are imposed on the spectral detection unit of the data.And the traditional early-warning technology is too slow to respond to water pollution events.When a pollution event occurs,whether it is traditional chemical measurement or water quality data transmission based on the Internet of Things technology,the model calculation of water quality parameters is still on the traditional local server.The speed is relatively slow for cloud servers.It takes up a lot of time in the round-trip data transmission and evaluation display,and it is impossible to efficiently make engineering response plans for water pollution incidents.At the same time,in the traditional water quality warning,most of the research on water pollution incidents It is independent and cannot make full use of the historical events that have occurred.The time required for accident warning is too long,and it cannot respond to the water quality situation quickly and quickly.Therefore,this study proposes a cloud platform-based water pollution warning method.Based on this,this thesis is jointly funded by the National Natural Science Foundation of China(61805029),Chongqing Social Enterprise and Livelihood Security Science and Technology Innovation Project(cstc2017shmsA1497)and Chongqing University of Technology Graduate Innovation Project(ycx20192053).Research on water pollution early warning methods based on cloud platform.The main research work of the paper is as follows:1)Design of UV-visible spectroscopy water quality detection experiment system.The american Ocean Optical DH2000 deuterium halogen lamp combined light source and american Ocean Optical Maya2000 Pro spectrometer were used to set up a UV-visible spectroscopy scientific-grade water quality detection system.Prepared experiments for different concentrations of potassium hydrogen phthalate solution standards,whilecollected distilled water,river water,domestic sewage and other water quality samples at the same time,and obtained the water quality UV-visible spectrum data of the samples.2)Research on the application of normalization algorithm based on EWMA-PCA in the standardization of water quality spectral data.Model transfer is of great significance to solve the inconsistency of the measurement signals caused by the different response functions of the samples and the instruments.The effective method to solve the model transfer is the standardization of the instrument or data.In view of the inconsistencies in the resolution,accuracy,and response range of UV-visible spectroscopy multi-parameter detection spectral detection units,it is difficult to compare test data between different instruments and to fit multi-parameter data.This paper proposes the use of EWMA-PCA normalized algorithm to realize the model transfer of UV-visible water quality spectrum on different instruments.The results of three sets of comparison experiments show that the algorithm can be well applied to different comparison spectrometers.After the water quality absorption spectrum data is standardized by the EWMA-PCA normalization algorithm,the correlation coefficient reaches 99.5765%,and the variance reaches0.0823 %,And the peak offset can be reduced to 0.0005%.3)Research on classification of water quality based on deep learning.In order to timely and effectively find occasional or deliberate water quality abnormal changes and protect the water environment,based on the existing water quality abnormality analysis,a water quality classification algorithm based on deep learning is proposed.This algorithm is aimed at the existing UV-visible spectroscopy water quality monitoring.The problems of spectral signals are susceptible to system noise interference,scattered solid particles scattering interference,and information redundancy,multicollinearity,etc.In addition,the existing water quality model has the problems of low accuracy and poor adaptability.A convolutional neural network is used to establish a water quality discriminant model and classify COD and NO3-N.4)Research on water pollution early warning method based on cloud platform.The establishment of a cloud-based water pollution early warning system can respond to the rapid changes in the current water environment,save computing time,and escort the environment and formulate pollution treatment policies in the next step.The cloud platform-based water pollution early warning method proposed in this paper provides a new idea for the spectral data standardization and water quality early warning of real-time online water quality monitoring systems.It has important guiding significance for the model transfer of instruments and the classification of water qualityconditions.-The practical application of visible spectrum water quality online monitoring project provides an efficient method.