The Study Of The Detection Technology Of Mineral Oil In Water Based On Three-dimensional Fluorescence Spectrum

With the continuous expansion of the scale of the land and offshore oil exploitation, and the gradually increased emissions of domestic sewage and industrial waste water from urban and rural area in our country, mineral oil contaminants in waste water are increasing and become more and more a threat to the safety of people and the environment. Therefore, rapid and accurate detection of the type and content of mineral oil in water is of great significance.The main methods of measuring the content of mineral oil in water are gravimetric method, turbidity, ultrasonic, gas liquid chromatography, spectrophotometry, fluorescence spectrometry and so on. But for trace contaminants in water, the measurement accuracy and techniques are very critical. In this paper, the fluorescence methods used at home and abroad are analyzed and studied, and a method of combination of the three-dimensional fluorescence spectra and characteristic parameter extraction based on fluorescence technique is proposed for the status of detection of mineral oil in water. Three-Dimensional Fluorescence Spectrum(TDFS) is a new method of fluorescence analysis. The three-dimensional data obtained by this method contains a rich amount of information, which increases the comprehensiveness and selectively of data processing.Firstly, the fluorescence analysis is discussed in this paper. The two-dimensional fluorescence properties of mineral oil are analyzed and the two-dimensional fluorescence spectroscopy of several mineral oil is achieved through experiments. The relationship between fluorescence intensity and the concentration of the test substance is studied. At the same time the limitations of the traditional two-dimensional fluorescence spectroscopy are pointed. Then the three-dimensional fluorescence spectroscopy mathematical principles, calculation methods and procedures are described.The fluorescence measurement system is designed, and the options for each component of the system are proposed. The pulsed xenon lamp is used as the excitation source and the blazed spectra-grating as the monochromator. The RL1024 SB type CCD(Charge-Coupled Devices) is designed as the photoelectric conversion device. The low loss quartz optical fiber bundle is used as optical signal transmission channel, and the dualchannel detection circuit which is made up by the preamplifier circuit, filter circuit and boxcar is designed. The weak fluorescence signal is extracted from background noise. The control unit of the system applies ARMLPC2210 single-chip microcomputer.The integrating-value of Three-Dimensional fluorescence spectrum volume is calculated by using Monte-Carlo integration method. The common methods, such as peak method, averaging method, and volumetric method are proposed so as to calculate the Three-Dimensional fluorescence spectrum volume. By means of the fluorescence spectrum experiment of a single element and mixing elements of mineral oil in water, the sensitive characteristic parameters such as the mean and the standard deviation of the Three-Dimensional fluorescence spectrum as well as the cumulative rate distribution of fluorescence intensity, the origin moments, the central moments, the orthocenter, Oval dense and so on are extracted.The working curve, sensitivity, and integration domain of the Monte-Carlo integration method of Three-Dimensional fluorescence spectrum volume are researched. Based on the wavelet analysis and the principle of Shannon entropy, the threshold de-noising algorithm of fluorescence signals is proposed. The wavelet entropy of de-noising analysis method is applied to detection signal of mineral oil in water, and the optimal wavelet de-noising function of fluorescence signals is confirmed by means of experiments. Comparing the traditional wavelet de-noising method and Shannon entropy of wavelet threshold de-noising method, two kinds of algorithm simulation data and conclusions are obtained. The signal-noise separation effect of Shannon entropy of wavelet threshold de-noising method to the fluorescence signals is verified.