本文選題：水質(zhì)遠(yuǎn)程在線監(jiān)測　切入點(diǎn)：曲線擬合　出處：《浙江理工大學(xué)》2017年碩士論文

【摘要】：目前以水質(zhì)分析儀為主的水質(zhì)分析設(shè)備雖備受青睞,但在實(shí)現(xiàn)遠(yuǎn)程在線監(jiān)測以及水樣(污水)毒性物質(zhì)成分與含量的分析方面仍缺失,同時(shí),監(jiān)測系統(tǒng)存在采集終端高并發(fā)的數(shù)據(jù)訪問、數(shù)據(jù)標(biāo)準(zhǔn)異構(gòu)導(dǎo)致難以實(shí)現(xiàn)數(shù)據(jù)流快速響應(yīng)和實(shí)時(shí)處理等問題。本文在研制出的水質(zhì)分析儀的基礎(chǔ)上,搭建了水質(zhì)遠(yuǎn)程在線監(jiān)測系統(tǒng),該系統(tǒng)以對水樣中毒性物質(zhì)成分和濃度的識(shí)別為核心目的,并根據(jù)數(shù)據(jù)流動(dòng)過程分別研究數(shù)據(jù)采集、數(shù)據(jù)處理和數(shù)據(jù)存儲(chǔ)相關(guān)設(shè)計(jì)和優(yōu)化技術(shù),實(shí)現(xiàn)對未知水樣的遠(yuǎn)程實(shí)時(shí)監(jiān)測的同時(shí)提高系統(tǒng)的通信性能與數(shù)據(jù)處理能力。本文主要完成了以下幾項(xiàng)工作:1、水質(zhì)遠(yuǎn)程在線監(jiān)測系統(tǒng)的搭建。針對目前水質(zhì)的遠(yuǎn)程在線監(jiān)測系統(tǒng)以及對水質(zhì)毒性物質(zhì)成分和濃度預(yù)測方面的短缺問題,設(shè)計(jì)并搭建了一種水質(zhì)遠(yuǎn)程在線監(jiān)測系統(tǒng),利用發(fā)光細(xì)菌發(fā)光原理,以明亮發(fā)光桿菌3變種作為毒性測試物種,實(shí)現(xiàn)了對水樣中毒性物質(zhì)成分和濃度的識(shí)別、水質(zhì)分析儀的遠(yuǎn)程管理與在線監(jiān)測。2、曲線擬合模型構(gòu)建與特征提取。針對使用常用的曲線擬合函數(shù)對反應(yīng)機(jī)理曲線進(jìn)行擬合時(shí),出現(xiàn)的變質(zhì)現(xiàn)象及擬合精度不高等問題,提出了一種基于改進(jìn)的B樣條曲線擬合算法,解決了在模型構(gòu)建時(shí)因追求擬合精度而違背毒性物質(zhì)對發(fā)光細(xì)菌抑制性作用等問題,并將擬合后的模型參數(shù)結(jié)合毒性物質(zhì)屬性設(shè)定為特征向量。3、毒性物質(zhì)成分和濃度的識(shí)別。由于數(shù)據(jù)的冗余性,對提取的特征向量使用PCA和LDA算法進(jìn)行降維處理,并結(jié)合BP神經(jīng)網(wǎng)絡(luò)對未知樣本中的毒性物質(zhì)成分進(jìn)行識(shí)別處理,后針對使用BP神經(jīng)網(wǎng)絡(luò)訓(xùn)練時(shí)所需要的迭代次數(shù)多、收斂速度慢以及容易出現(xiàn)在未達(dá)到訓(xùn)練目標(biāo)時(shí)訓(xùn)練終止等問題,提出了基于改進(jìn)的BP神經(jīng)網(wǎng)絡(luò)算法對毒性物質(zhì)成分和濃度進(jìn)行識(shí)別,有效地改善了算法的性能,且LDA與改進(jìn)后的BP網(wǎng)絡(luò)模型相結(jié)合對毒性物質(zhì)成分的識(shí)別正確率達(dá)到了100%,濃度識(shí)別率達(dá)到了92%以上。4、對水質(zhì)遠(yuǎn)程在線監(jiān)測系統(tǒng)的優(yōu)化與研究。針對系統(tǒng)實(shí)際應(yīng)用中遇到的高并發(fā)訪問請求、數(shù)據(jù)的實(shí)時(shí)性能、數(shù)據(jù)傳輸效率以及數(shù)據(jù)包的封裝和解析速率低等問題,分別對數(shù)據(jù)采集、數(shù)據(jù)處理以及數(shù)據(jù)存儲(chǔ)三個(gè)方面進(jìn)行優(yōu)化。面對高速率數(shù)據(jù)訪問請求時(shí),在IOCP模型的基礎(chǔ)上,提出了基于對象池模式的自適應(yīng)線程池技術(shù),有效的解決了對共享資源的并發(fā)訪問效率低的問題,提高了系統(tǒng)的通信效率。針對數(shù)據(jù)傳輸效率以及數(shù)據(jù)包的封裝和解析速率,提出了基于JSON和TLV的消息格式優(yōu)化。最后針對數(shù)據(jù)的實(shí)時(shí)性和系統(tǒng)的使用率等問題,提出了高效數(shù)據(jù)流處理算法,對數(shù)據(jù)庫管理技術(shù)進(jìn)行改進(jìn)。5、將以上研究成果用VC++語言實(shí)現(xiàn),按一定的邏輯集成為數(shù)據(jù)處理模塊,加入到水質(zhì)遠(yuǎn)程在線監(jiān)測系統(tǒng)的服務(wù)器端,使采集到的數(shù)據(jù)得到了利用,最后通過系統(tǒng)的實(shí)現(xiàn)和實(shí)際應(yīng)用驗(yàn)證了本文提出的曲線擬合構(gòu)建、毒性物質(zhì)成分和識(shí)別的方法。研究結(jié)果表明,本文提出的基于改進(jìn)的B樣條曲線擬合方法彌補(bǔ)了常見擬合函數(shù)的擬合精度不佳和“變質(zhì)失性”問題,提高了曲線擬合效果,為后續(xù)的特征提取工作提供了有利的支撐;基于改進(jìn)的BP神經(jīng)網(wǎng)絡(luò)算法,彌補(bǔ)了傳統(tǒng)BP神經(jīng)網(wǎng)絡(luò)的不足,提高了毒性物質(zhì)成分和濃度的識(shí)別正確率;基于完成端口模型的網(wǎng)絡(luò)通信性能的優(yōu)化,解決了在實(shí)際應(yīng)用中遇到的并發(fā)處理能力不強(qiáng)、數(shù)據(jù)傳輸效率低以及數(shù)據(jù)包的封裝和解析速率低等問題,使系統(tǒng)的穩(wěn)定性、性能得到了改善。
[Abstract]:Water quality analysis equipment at present water quality analyzer mainly is favored, but in the realization of remote online monitoring and water (sewage) analysis of ingredients and contents of toxic substances are still missing, at the same time, there are high concurrent access to the data acquisition terminal monitoring system, which leads to the problem of heterogeneous data standards difficult to achieve fast response and real-time data stream processing. Based on the water quality analyzer developed on set up a remote on-line monitoring system, the system to identify the constituents and concentration of water poisoning is the core purpose, and according to the data flow process were studied in data acquisition, data processing and data storage design and optimization technology to improve system's communication performance and the data processing ability of unknown samples in the remote real-time monitoring at the same time. This paper mainly completes the following work: 1, remote online monitoring of water quality To build the test system. At present, remote online monitoring system of water quality and water quality of toxic substance composition and concentration prediction of shortage, designed and built a water quality remote monitoring system, using the principle of luminescent bacteria, to Photobacterium phosphoreum toxicity test as 3 varieties of species, the identification of the material composition and concentration for water poisoning, remote management and online monitoring of.2 water quality analyzer, construction and feature extraction of curve fitting model. According to the curve fitting function is used to fit the curve of reaction mechanism, metamorphic phenomenon and the fitting accuracy is not high, we propose an improved B algorithm based on spline curve fitting, solving the contrary problem of toxic substances on the luminescent bacteria inhibition effect due to the pursuit of the fitting accuracy in the model construction, and combining the model parameters after fitting Toxic property is set to feature vector.3, identification of material composition and concentration of toxicity. Because of the redundancy of data, the dimensionality reduction process using PCA and LDA algorithm for feature extraction, and combining with the BP neural network recognition processing of toxic substances in the composition of unknown samples, the number of iterations for the use of BP neural network training when needed, slow convergence and easy to appear in the training target training does not reach the termination problem, put forward the improved BP neural network algorithm based on the composition and concentration of toxic substances are identified, effectively improve the performance of the algorithm, and the BP LDA network model and the improved combination of identification of toxic substances composition of the correct rate reached 100%, the concentration of recognition rate reached more than 92%.4, optimization and Research on remote monitoring system of water quality. The high concurrency encountered in practical application systems The access request, the real-time performance of data, the efficiency of data transmission and data packet encapsulation and resolution rate low, respectively on the three aspects of data acquisition, data processing and data storage optimization. With high speed data access request, based on the IOCP model, proposed an adaptive thread pool technique based on Object Pool Pattern and effectively solves the problem of low efficiency of concurrent access to shared resources, improve the efficiency of communication system. The efficiency of data transmission and data packet encapsulation and resolution rate, put forward the optimization of message format JSON and based on TLV. Finally according to the real-time data and system usage, put forward the flow efficient data processing algorithm, improved.5 database management technology, the research results with the VC++ language, according to certain logic integrated data processing module is added to the water. Quality of remote online monitoring system server, the collected data are utilized, and finally through the system implementation and practical application verify the curve fitting of this construction method, toxic substance composition and identification. The results show that the proposed method improved the B spline curve fitting based on offset fitting accuracy the common fitting function is not good and the "lost" problem of deterioration, improve the effect of curve fitting, provide a favorable support for the subsequent feature extraction; improved BP neural network algorithm based on traditional BP neural network, improve the recognition of material composition and concentration of the toxicity of the correct rate of network performance optimization; communication completion port model based on solving the concurrent processing capability encountered in the practical application is not strong, low efficiency of data transmission and data packet encapsulation and resolution rate etc. The problem is that the stability of the system and the performance of the system have been improved.

【學(xué)位授予單位】：浙江理工大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：R123.1;TP274

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 黃燦克;劉婷婷;湯曉畏;;發(fā)光細(xì)菌毒性法在飲用水水質(zhì)評(píng)估與預(yù)警中的應(yīng)用[J];環(huán)境監(jiān)控與預(yù)警;2015年03期

2 徐建英;趙春桃;魏東斌;;生物毒性檢測在水質(zhì)安全評(píng)價(jià)中的應(yīng)用[J];環(huán)境科學(xué);2014年10期

3 余朋;;網(wǎng)絡(luò)通信協(xié)議的分析與實(shí)現(xiàn)[J];電腦編程技巧與維護(hù);2014年14期

4 王祖麟;高萌;肖翔群;;便攜式遠(yuǎn)程深水原位水質(zhì)在線監(jiān)測系統(tǒng)[J];電子技術(shù);2013年07期

5 杜兵;胡愛蘭;張子郡;;基于IOCP機(jī)制的負(fù)控管理系統(tǒng)研究[J];化工自動(dòng)化及儀表;2013年05期

6 王茜劏;黃志文;劉凱;李文江;閻吉祥;;基于主成分分析和人工神經(jīng)網(wǎng)絡(luò)的激光誘導(dǎo)擊穿光譜塑料分類識(shí)別方法研究[J];光譜學(xué)與光譜分析;2012年12期

7 何桂華;田松坡;譚劍亮;許子良;葉千均;;新型水質(zhì)毒性分析儀研制[J];電子測量與儀器學(xué)報(bào);2012年10期

8 魯金濤;李夕兵;宮鳳強(qiáng);王希然;柳皎;;基于主成分分析與Fisher判別分析法的礦井突水水源識(shí)別方法[J];中國安全科學(xué)學(xué)報(bào);2012年07期

9 亓開元;趙卓峰;房俊;馬強(qiáng);;針對高速數(shù)據(jù)流的大規(guī)模數(shù)據(jù)實(shí)時(shí)處理方法[J];計(jì)算機(jī)學(xué)報(bào);2012年03期

10 JUNG D Y;LEE S M;王洪梅;KIM J H;LEE S H;;Fault detection method with PCA and LDA and its application to induction motor[J];Journal of Central South University of Technology;2010年06期

相關(guān)碩士學(xué)位論文 前10條

1 周鵬飛;基于改進(jìn)的模糊BP神經(jīng)網(wǎng)絡(luò)的圖像分割方法研究[D];太原理工大學(xué);2014年

2 邊國興;基于IOCP和多線程技術(shù)的通訊中間件設(shè)計(jì)與實(shí)現(xiàn)[D];電子科技大學(xué);2014年

3 李一君;水質(zhì)在線監(jiān)測系統(tǒng)的研究與設(shè)計(jì)[D];南昌大學(xué);2012年

4 樹愛兵;基于IOCP的交通信號(hào)控制通信服務(wù)器研究[D];上海交通大學(xué);2012年

5 路林;即時(shí)通信協(xié)議的特征與通聯(lián)關(guān)系分析[D];解放軍信息工程大學(xué);2012年

6 于濤;BP網(wǎng)絡(luò)自適應(yīng)學(xué)習(xí)率算法分析[D];大連理工大學(xué);2011年

7 張明亮;基于嵌入式技術(shù)的水質(zhì)分析儀關(guān)鍵技術(shù)的研究[D];浙江理工大學(xué);2011年

8 呂先濤;基于BP網(wǎng)絡(luò)混合氣體濃度識(shí)別的研究[D];武漢理工大學(xué);2010年

9 張波;基于嵌入式水質(zhì)綜合生物毒性在線自動(dòng)分析儀控制器關(guān)鍵技術(shù)的研究[D];浙江理工大學(xué);2010年

10 劉俊超;基于正則表達(dá)式的應(yīng)用層協(xié)議識(shí)別技術(shù)研究[D];國防科學(xué)技術(shù)大學(xué);2008年

，

本文編號(hào)：1722202