船舶排放污染物智能监测系统研究

发布时间：2018-09-02 07:43

【摘要】：船舶含油污水以及废气中硫化物的排放导致海洋环境污染的加剧,受到了国际社会的广泛关注。船舶排放污染物的有效监测对海洋环境的保护起着至关重要的作用。传统船用油分浓度计在实际检测中易受气泡等干扰因素的影响,存在非线性偏差影响检测精度。而对于船舶废气中硫化物等污染物的在线监测,目前市场上还没有一套完全成熟的系统实际应用于船舶废气的在线监测。遵守相关公约环保减排,完全依靠船员和船东的自觉性。在“智能船舶”的发展新趋势下,面对现实应用的需求,本文研究设计了一种船舶排放污染物智能监测系统。系统主要包括船用油分浓度计原理样机和船舶废气中二氧化硫浓度检测原理样机以及上位机监测软件三部分。本文以光学检测技术和软测量法相结合的研究为背景,深入开展了船舶排放污染物智能监测系统的研究。论文的主要研究工作是:1.船用油分浓度计原理样机的研制。(1)船用油分浓度计原理样机的硬件设计。主要包括传感器光电结构设计、硬件电路设计以及信号采集处理模块的设计。(2)基于最小二乘支持向量机(Least Squares Support Vector Machine,LS-SVM)的船用油分浓度检测模型研究。针对基于浊度法的传统船用油分浓度检测中存在气泡等干扰因素导致非线性偏差,同时超过一定浓度的检测范围,需要进行非线性补偿计算的问题,采用LS-SVM建立油分浓度的检测模型。LS-SVM在解决小样本统计、非线性建模方面具有很好的应用。实验结果表明,其可用于船用油分浓度计原理样机的研制。(3)粒子群优化(Particle Swarm Optimization,PSO)算法对基于LS-SVM的船用油分浓度检测模型参数的优化研究。针对基于LS-SVM的船用油分浓度检测模型参数人为选择的盲目性,影响模型预测精度的问题。采用PSO算法优化LS-SVM油分浓度检测模型的参数。以均方根误差(RMSE)和平均相对误差(MRE)作为评价标准,对比LS-SVM船用油分浓度预测模型以及采用最小二乘拟合的结果。采用粒子群算法优化最小二乘支持向量机模型参数的PSO-LS-SVM油分浓度检测模型精度更高,且有效避免了由于模型参数选择不当,导致模型泛化性能降低的问题。适用于船用油分浓度计原理样机的研制。2.基于小波分析的二氧化硫浓度检测信号去噪算法研究。针对实际检测中二氧化硫对红外光的吸收很弱,检测信号容易被淹没在噪声中,在经过放大滤波后,信号仍不免受到放大器、外界环境、辐射源等带来的噪声影响,利用小波去噪算法的时域和频域局域性、检测信号奇异性和突变结构的优点,对含噪信号进行小波阈值去噪。3.开展了基于Lab VIEW软件平台的船舶排放污染物智能监测系统上位机监测软件的设计。
[Abstract]:The pollution of marine environment caused by the discharge of oil-bearing sewage and sulfide from the waste gas of ships has attracted extensive attention of the international community. Effective monitoring of pollutants discharged from ships plays an important role in the protection of marine environment. Nonlinear deviation affects the detection accuracy. However, there is not a complete set of mature system on the market for on-line monitoring of sulfide and other pollutants in ship exhaust gas. In order to meet the needs of practical application, this paper studies and designs an intelligent monitoring system for marine pollutants. The system mainly includes three parts: the principle prototype of marine oil concentration meter, the principle prototype of sulfur dioxide concentration detection in ship exhaust gas and the monitoring software of upper computer. The main research work of this paper is as follows: (1) The development of marine oil concentration meter prototype. (1) The hardware design of marine oil concentration meter prototype. It mainly includes the design of sensor photoelectric structure, hardware circuit and signal acquisition and processing module. (2) Research on the detection model of marine oil concentration based on Least Squares Support Vector Machine (LS-SVM). For the non-linear deviation caused by bubbles and other interference factors in the traditional detection of marine oil concentration based on turbidity method, which exceeds the detection range of a certain concentration, non-linear compensation is needed. LS-SVM has good application in solving small sample statistics and nonlinear modeling. The experimental results show that it can be used in the development of marine oil concentration meter prototype. (3) Particle Swarm Optimization (PSO) algorithm for LS-SVM-based marine oil content. Concentration detection model parameters optimization research. Aiming at the blindness of parameter selection of marine oil concentration detection model based on LS-SVM, which affects the prediction accuracy of the model, PSO algorithm is used to optimize the parameters of LS-SVM oil concentration detection model. The PSO-LS-SVM oil concentration detection model optimized by particle swarm optimization algorithm has higher accuracy and avoids the problem of model generalization performance degrading due to improper model parameters selection. 2. Research on the denoising algorithm of sulfur dioxide concentration detection signal based on wavelet analysis. Aiming at the fact that the absorption of sulfur dioxide to infrared light is very weak, the detection signal is easily submerged in noise. After amplification and filtering, the signal is still unavoidable from noise caused by amplifier, external environment, radiation source and so on. The wavelet threshold denoising method is used to denoise the noisy signal by using the time domain and frequency domain locality of the wavelet denoising algorithm and the advantages of detecting singularity and abrupt structure of the signal. 3. The upper computer monitoring software of the ship emission intelligent monitoring system based on Lab VIEW software platform is designed.
【学位授予单位】：江苏科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP274

【参考文献】