Water is the source of life and is closely related to human life and production.Carrying out water quality monitoring is a long-term and lasting task.Traditional water quality monitoring is mainly based on chemical method,which has the disadvantages of long measurement cycle,complicated operation,secondary contamination of reagents,and poor real-time performance,etc.With the development of the new concept of water quality detection based on direct spectrum,it has laid a foundation for rapid,efficient,in-situ,green,real-time and online water quality monitoring,which forms complementary advantages with traditional water quality monitoring.In response to the cross-sectional water quality monitoring needs,the domestic and foreign direct spectroscopic water quality detection technology and its current status are studied in this research.And an immersed water quality automatic monitoring system based on spectroscopy is developed,including immersed water quality automatic monitoring equipment and remote platform system.The main research work includes the following aspects:(1)The principle of ultraviolet-visible full-spectrum water quality detection is studied.In addition,the general framework of an immersed water quality automatic monitoring system based on spectroscopy is proposed,including hardware monitoring equipment and software platform system.Hardware and software are coordinated to build an all-weather,efficient,green and real-time online monitoring mode for water quality.(2)According to the detection requirements of spectroscopy,an immersed water quality automatic monitoring equipment based on spectroscopy is designed,including the mechanical part and the electronic control part.In order to meet the requirements of in-situ detection of spectroscopy,the overall composition of the mechanical part of the immersed water quality automatic monitoring equipment is determined.Besides,the measurement framework,outer cylinder wall,open sink,mirror cleaning device,and water solar power supply device are designed separately,and finite element software is used to carry out lightweight design.According to the action flow of each execution unit in the detection process,the hardware circuit and control program of the electronic control module are designed.(3)A Keras-based artificial neural network is proposed for water quality parameter detection modeling.Spectral data is obtained through the spectrum test of mixed standard solution(prepared by orthogonal mixing of turbidity(TSS),chemical oxygen demand(COD)and total nitrogen(TN)standard solutions).The coarse-grained response bands of TSS,COD and TN are qualitatively analyzed from the absorbance curve and additive absorbance curve;and the influence of TSS,COD and TN on the corresponding bands is quantitatively analyzed by range method with the integral area as the result of orthogonal spectrum experiment.Therefore,the optimal response band of TSS,COD and TN is confirmed.According to the spectral response band,the TSS,COD,and TN artificial neural network index detection models are established.The results show that the TSS,COD,and TN models have good detection accuracy,and the detection accuracy is respectively controlled within 10%,15%,and 20%.(4)A remote platform prototype system for automatic water quality monitoring is designed and developed based on B/S(Browser/Server)architecture.In order to promote the further improvement of the construction of the Internet of Things for automatic water quality monitoring,the real-time online water quality monitoring is realized combined with immersed monitoring equipments,artificial neural network water quality parameter detection models and communication networks. |