Research On Curve Fitting Prediction Model And Algorithm In Remote Water Quality Monitoring System

At present,water quality analysis equipment is favored,like water quality analyzer,but it is still short at the analysis of remote on-line monitoring and the composition and content of water samples(sewage)toxic substances.Also,there is a high level of data access in the monitoring system,and the heterogeneity of data standard makes it difficult to realize the fast response and real-time processing of data flow.Based on the developed water quality analyzer,this paper builds a remote monitoring system of water quality,which aims at identifying the composition and concentration of toxic substances in water samples,and according to the data flow process,the data acquisition,data processing and data storage related design and optimization techniques are studied respectively to realize the remote real-time monitoring of unknown water samples while improving the communication performance and data processing capability of the system.This paper mainly completed the following work:1,The remote monitoring system for water quality.In order to solve the shortcomings of the current system and improve the accuracy of the prediction,a remote monitoring system of water quality is designed.According to the luminous bacterium luminescence principle,the bright luminous bacteria 3 variants are used as the toxicity test species to achieve the water sample of toxic substances in the composition and concentration of identification,water quality analyzer remote management and on-line monitoring.2,Curve fitting model construction and feature extraction.In this paper,an improved B-spline curve fitting algorithm is proposed to solve the problem that the metamorphic phenomena and the fitting precision are not high when the curve is fitted by the commonly used curve fitting function.The problem of suppressing the inhibitory effect of the toxic substances on the luminous bacteria is solved by the pursuit of the fitting precision,and the fitted model parameters combined with the toxic substance properties are set as the eigenvector.3,Identification of toxic constituents and concentrations.Due to the redundancy of the data,the PCA and LDA algorithms are used to reduce the dimensionality of the extracted eigenvectors,and the components of the toxic substances in the unknown samples are identified by BP neural network.Then,aiming at the problem that the number of iterations required for training using BP neural network is high,the convergence rate is slow and it is easy to occur when the training target is not reached.The improved BP neural network algorithm is used to identify the toxic substance composition and concentration,the improved BP neural network algorithm is used to identify the toxic component and concentration,and the performance of the algorithm is improved effectively.And the combination of LDA and improved BP network model to the toxicity of the composition of the correct rate of 100%,the concentration of recognition rate of 92% or more.4,Optimization and research on remote online monitoring system for water quality.The data acquisition,data processing and data storage are optimized in view of the high concurrent access request,the real-time performance of data,the efficiency of data transmission and the low packet and parsing rate encountered in the practical application of the system.In the face of high-speed data access request,based on the IOCP model,an adaptive thread pool technology based on object pool model is proposed,which effectively solves the problem of low efficiency of concurrent access to shared resources and improves the communication efficiency of the system The Aiming at the efficiency of data transmission and the encapsulation and resolution rate of data packets,the message format optimization based on JSON and TLV is proposed.Finally,aiming at the real-time data and the utilization rate of the system,an efficient data stream processing algorithm is proposed to improve the database management technology.5,The above research results with VC ++ language,according to a certain logic into a data processing module,added to the water quality remote monitoring system,so that the collected data has been used.Finally,the method of curve fitting,the composition of toxic substances and the method of identification are presented through the realization and practical application of the system.The results show that the improved B-spline curve fitting method proposed in this paper solves the problem of poor fitting precision and "metamorphic failure" in common fitting process,which provides a favorable support for the subsequent feature extraction.Based on the improved BP neural network algorithm,it can make up the deficiency of the traditional BP neural network and improve the correctness of the identification and concentration of toxic substances.Based on the optimization of the network communication performance of the port model,it solves the problem of low concurrency processing capability,low data transmission efficiency and low packet and parsing rate encountered in practical application,so that the stability of the system,performance has been improved.