CN105092523A

CN105092523A - Scheme based on near infrared technology and used for detecting phosphorous in water

Info

Publication number: CN105092523A
Application number: CN201510525993.9A
Authority: CN
Inventors: 尹燕燕; 张世勤; 刘燕卿
Original assignee: Jiangnan University
Current assignee: Jiangnan University
Priority date: 2015-08-25
Filing date: 2015-08-25
Publication date: 2015-11-25

Abstract

Disclosed is a scheme based on near infrared technology and used for detecting phosphorous in water. The scheme includes: utilizing near infrared spectroscopy to measure 0mg/L-50mg/L of phosphorous solutions different in concentration by utilizing NIR equipment to acquire spectra, and storing the spectra as CSV data; utilizing TQ analyst for primary analysis; utilizing a derivative algorithm (Derivatives norris gap), a multiplicative scatter correction (Multiplicative scatter correction, MSC) algorithm, a smoothing (Smoothing) algorithm and a standard normal variate transformation (Standard normal variate transformation, SNV) method to pre-process near infrared spectrum data respectively; utilizing MATLAB for simulation, selecting a simulation model PLSR, utilizing a concentration gradient classifying method for simulation, and building a mathematic model; acquiring simulation results, optimizing the same, and integrating to acquire a final result. By the scheme, quick, environment-friendly, nondestructive and online detection of the phosphorous solutions different in concentration can be realized.

Description

A kind of detection based on phosphorus in the water quality of near infrared technology

Technical field

The present invention relates to a kind of detection based on phosphorus in the water quality of near infrared technology, belong to chemical detection analysis field.

Background technology

423, whole nation trunk river, 62 emphasis lakes (reservoir) have altogether 968 states's control surface water monitoring section (some position), water quality monitoring is carried out to these sections, I, II, III, IV, V, bad V class water quality section accounts for 3.4%, 30.4%, 29.3%, 20.9%, 6.8%, 9.2% respectively, wherein one of main contamination index is total phosphorus.

In spring, summer and autumn, the oceanic area being inferior to the 4th class sea water quality standard in the full marine site of China is respectively 52280 sq-kms, 41140 sq-kms and 57360 sq-kms, wherein in the marine site of national offshore, 301 state's controlling/monitoring points detect data display, one, two, three, four and the seawater of bad four classes account for 28.6%, 38.2%, 7.0%, 7.6%, 18.6% respectively, but main contamination index is inorganic nitrogen and reactive phosphate.

Tradition unification of the motherland assay method has ammonium molybdate spectrophotometric method and connected neighbor set.Classic method can use chemical reagent in mensuration process, damages sample itself, meanwhile, also can bring error, and sample reagent cannot reduce after mensuration, more cannot revert in original system.

Summary of the invention

The object of the invention is to overcome the deficiency existed in conventional art, provide a kind of detection based on phosphorus in the water quality of near infrared technology, it is easy to operate, and detection speed is fast, and environmental protection, does not have destructiveness to sample, and can on-line checkingi.

According to technical scheme provided by the invention, be divided into following steps:

The first step, the preparation of solution.

Second step, utilizes NIR checkout equipment to detect, and each detection all will carry out cleaning-drying to instrument probe.Nearly infared spectrum saves as CSV data.

3rd step, pre-service.Derivative algorithm (Derivativesnorrisgap), multiplicative scatter correction algorithm (Multiplicativescattercorrection, MSC), level and smooth (Smoothing) algorithm and standard normal variable conversion (Standardnormalvariatetrans-formation, SNV).

4th step, utilizes MATLAB to emulate.Choose realistic model PLSR, and utilize concentration gradient sorting technique, founding mathematical models.Draw simulation result and optimize, integrating and draw net result.

5th step, enter contrast to MATLAB data, utilize index RMSEC, Rc, RMSEP and Rp compare the model set up in different pretreatments situation respectively, filter out optimal set.

The method can realize canonical correlation analysis, multiple linear regression and principal component analysis between two groups of variablees simultaneously.Therefore, the data processing of the multivariate statistics advantage that has it exclusive.In PLS, its ultimate principle and algorithmic approach are mainly independent variable with X, and Y is that dependent variable makes a matrix.Therefore, its model is established as a bilinear model.So, in original data processing process, using wavelength and 102 groups as an independently X matrix-block (102 × 1557), using the actual value of 102 groups as Y matrix-block (102 × 1).Can external relations when Modling model, two independent matrix blocks and internal relations, the relation between two matrix-blocks forms final mathematical model.By the linear regression relation existed between independent variable and the latent variable of dependent variable, X matrix-block and the direct mutual relationship of Y matrix-block are described, carry out mathematical modeling with this.

Accompanying drawing explanation

The arrangement of Fig. 1 Fourier transform infrared spectrometer device and operating diagram;

Fig. 2 is based on the testing process process flow diagram of phosphorus near infrared technology water quality.

Embodiment

Below in conjunction with concrete accompanying drawing, the present invention will be further described.

As Fig. 1 Fourier transform infrared spectrometer device basic composition comprises following five parts:

1. light emission system is analyzed.This system is made up of parts such as light source, beam splitter, samples, collects the analysis light carrying sample message.

2. interferometer.This instrument is core component, and its quality directly has influence on the treatment effect of whole instrument.The interferometer that tradition uses, michelson interferometer, also has the interferometer improved in addition.Their principle of work is all that the light beam that analysis optical transmitting set is sent is divided into two bundles, forms optical path difference, thus produces space or the analysis light expressed by time domain, namely interference light.

Michelson interferometer system is made up of the level crossing that two are mutually an angle of 90 degrees, index glass horizontal glass.Also have optical beam-splitter, light source and detector in addition.Index glass keeps vertical angle with horizontal glass all the time in the process of movement.Generalized case is in order to reduce vibration, and index glass is connected mobile by air bearing, beam splitter and light source are miter angle.The monochromatic light of 50% can pass through, and other 50% is reflected.Reflected light A and transmitted light B.Light beam A meeting vertical irradiation is on horizontal glass, and be then reflected back toward beam splitter, 50% is reflected back toward light source, and 50% injects detector through optical splitter.B light beam injects detector with same process together with A light beam.A light beam and B light beam here merge, and form the coherent light with interference light character.Utilize index glass constantly to change position, finally can obtain the interference light intensity of different optical path difference.

3. detector.Detect interference light.

4. sampling system.Utilize digital to analog converter that interference light is converted to digital signal, and utilize computer system to carry out sampling statistics.

5. computer processing system.Use computer system processor sample interference light function and light source interference light function, then through Fourier transform mathematics manipulation, interferogram is reduced into spectrogram, use display display can obtain near infrared light spectrogram.

The testing process process flow diagram of phosphorus in Fig. 2 near infrared water quality

1. preparing standard solution.Utilize the standard solution preparation 1mg/L-50mg/L of phosphorus, calibration is the phosphorus solution of 1mg/L, accomplishes, with joining, to reduce the solution impact that standing time, long environment produced it.

2. near infrared spectrum is measured.To step 1. in phosphorus solution carry out near infrared ray.Often organize sample and measure twice, reduce error.And CSV formatted data is saved as to the 102 groups of data obtained.

3. TQAnalyst analyzes.Initial analysis is carried out to the 102 groups of data matrixes obtained, obtains Preliminary Analysis Results, according to analysis result determination preprocess method and modeling method.

4. TQAnalyst makes matched curve.According to matched curve remove out-of-bounds point, without not good enough.

5. pre-service.Method is as follows: derivative algorithm (Derivativesnorrisgap), multiplicative scatter correction algorithm (Multiplicativescattercorrection, MSC), level and smooth (Smoothing) algorithm and standard normal variable conversion (Standardnormalvariatetrans-formation, SNV).

6. MATLAB simulation modeling.Modeling and simulating is carried out to pre-processed results, adopts PLS modeling method.

7. comparative result.Enter contrast to MATLAB data, utilize index RMSEC, Rc, RMSEP and Rp compare the model set up in different pretreatments situation respectively, filter out optimal set.

8. optimal case is obtained.By the optimal set filtered out, draw optimal case.

9. optimization model is utilized to detect total phosphorus content in unknown solution.According to optimal case and realistic model, unknown solution is substituted in model, the content of phosphorus in unknown solution is measured.

Claims

1., based on a detection for phosphorus in the water quality of near infrared technology, it is characterized in that, following steps:

(1) preparation of solution.

(2) utilize NIR checkout equipment to detect, obtain near infrared collection of illustrative plates.

(3) simple process is carried out to raw data, and pre-service.

(4) utilize MATLAB to emulate, choose realistic model PLSR, and utilize concentration gradient sorting technique, founding mathematical models.Draw simulation result and optimize, integrating and draw net result.

(5) enter contrast to MATLAB data, utilize index RMSEC, Rc, RMSEP and Rp compare the model set up in different pretreatments situation respectively, filter out optimal set, thus obtain optimization model.

2. solution preparation according to claim 1, is characterized in that 25% standard ammonia nitrogen solution to dilute, is mixed with the ammonia nitrogen solution that 0mg/L-50mg/L calibration is 1mg/L.

3. the NIR of utilization checkout equipment according to claim 1 detects, and obtains near infrared collection of illustrative plates.It is characterized in that each detection all will carry out cleaning-drying to instrument probe.Finally obtain the near infrared collection of illustrative plates of variable concentrations and save as CSV data.

4. according to claim 1 simple process is carried out to raw data.It is characterized in that utilizing the process of TQAnalys software, compare from image and draw out-of-bounds point, and remove out-of-bounds and error point.

5. TQAnalys software according to claim 4 process, is characterized in that, choose modeling method, partial least square method, process total data, draws primary out-of-bounds point, and removes.

6. pre-service according to claim 1, it is characterized in that utilizing derivative algorithm (Derivativesnorrisgap), multiplicative scatter correction algorithm (Multiplicativescattercorrection, MSC), level and smooth (Smoothing) algorithm and standard normal variable conversion (Standardnormalvariatetrans-formation, SNV) method carry out pre-service near infrared spectrum data respectively.

7. the MATLAB of utilization according to claim 1 emulates, choose realistic model PLSR, it is characterized in that utilizing MATLAB software to carry out simulation modeling to preprocessed data, PLSR model is chosen in emulation, due to the process that preprocessing process is only to near-infrared wavelength, the Y-axis (102 × 1557) namely in MATLAB data modeling.In addition, need to set up an X-axis (102 × 1). in this experiment, one has 102 groups of data groups.Therefore, the foundation of X-axis needs 102 x, and corresponding with actual value.

8. enter contrast to MATLAB data according to claim 1, utilize index RMSEC, Rc, RMSEP and Rp compare the model set up in different pretreatments situation respectively, it is characterized in that Rc, and Rp is closest to 1, and meanwhile, RMSEC, RMSEP numerical value is less.Optimization model can be obtained.