CN109270017A - Automatic water quality monitoring system and monitoring method under a kind of multifunctional water based on improved RBFNN algorithm - Google Patents
Automatic water quality monitoring system and monitoring method under a kind of multifunctional water based on improved RBFNN algorithm Download PDFInfo
- Publication number
- CN109270017A CN109270017A CN201811138411.1A CN201811138411A CN109270017A CN 109270017 A CN109270017 A CN 109270017A CN 201811138411 A CN201811138411 A CN 201811138411A CN 109270017 A CN109270017 A CN 109270017A
- Authority
- CN
- China
- Prior art keywords
- water
- water quality
- rbfnn
- sample
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 224
- 238000000034 method Methods 0.000 title claims abstract description 59
- 238000012544 monitoring process Methods 0.000 title claims abstract description 58
- 238000012549 training Methods 0.000 claims abstract description 59
- 238000012360 testing method Methods 0.000 claims abstract description 57
- 238000012545 processing Methods 0.000 claims abstract description 32
- 239000003153 chemical reaction reagent Substances 0.000 claims abstract description 9
- 238000005259 measurement Methods 0.000 claims description 69
- IJGRMHOSHXDMSA-UHFFFAOYSA-N Atomic nitrogen Chemical compound N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 claims description 22
- 230000006870 function Effects 0.000 claims description 21
- 238000004448 titration Methods 0.000 claims description 20
- XUIMIQQOPSSXEZ-UHFFFAOYSA-N Silicon Chemical compound [Si] XUIMIQQOPSSXEZ-UHFFFAOYSA-N 0.000 claims description 16
- 229910052710 silicon Inorganic materials 0.000 claims description 16
- 239000010703 silicon Substances 0.000 claims description 16
- 238000002798 spectrophotometry method Methods 0.000 claims description 13
- 238000011478 gradient descent method Methods 0.000 claims description 12
- 238000004458 analytical method Methods 0.000 claims description 11
- 229910052757 nitrogen Inorganic materials 0.000 claims description 11
- 238000003911 water pollution Methods 0.000 claims description 10
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 238000013461 design Methods 0.000 claims description 8
- 230000003287 optical effect Effects 0.000 claims description 8
- 230000006872 improvement Effects 0.000 claims description 7
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 claims description 6
- 239000013078 crystal Substances 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 5
- 239000003795 chemical substances by application Substances 0.000 claims description 5
- 238000005457 optimization Methods 0.000 claims description 5
- 230000010355 oscillation Effects 0.000 claims description 5
- 230000008901 benefit Effects 0.000 claims description 3
- 230000003750 conditioning effect Effects 0.000 claims description 3
- 238000013480 data collection Methods 0.000 claims description 3
- 230000007423 decrease Effects 0.000 claims description 3
- 238000009826 distribution Methods 0.000 claims description 3
- 229910052742 iron Inorganic materials 0.000 claims description 3
- 239000007788 liquid Substances 0.000 claims description 3
- NJPPVKZQTLUDBO-UHFFFAOYSA-N novaluron Chemical compound C1=C(Cl)C(OC(F)(F)C(OC(F)(F)F)F)=CC=C1NC(=O)NC(=O)C1=C(F)C=CC=C1F NJPPVKZQTLUDBO-UHFFFAOYSA-N 0.000 claims description 3
- 230000037452 priming Effects 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 238000007789 sealing Methods 0.000 claims description 3
- 229910001220 stainless steel Inorganic materials 0.000 claims description 3
- 239000010935 stainless steel Substances 0.000 claims description 3
- 238000005553 drilling Methods 0.000 claims description 2
- HUTDUHSNJYTCAR-UHFFFAOYSA-N ancymidol Chemical compound C1=CC(OC)=CC=C1C(O)(C=1C=NC=NC=1)C1CC1 HUTDUHSNJYTCAR-UHFFFAOYSA-N 0.000 claims 1
- 238000003780 insertion Methods 0.000 claims 1
- 230000037431 insertion Effects 0.000 claims 1
- 238000013386 optimize process Methods 0.000 claims 1
- 230000005622 photoelectricity Effects 0.000 claims 1
- 239000003643 water by type Substances 0.000 claims 1
- 238000007405 data analysis Methods 0.000 abstract description 2
- 230000003321 amplification Effects 0.000 description 8
- 238000003199 nucleic acid amplification method Methods 0.000 description 8
- KMUONIBRACKNSN-UHFFFAOYSA-N potassium dichromate Chemical compound [K+].[K+].[O-][Cr](=O)(=O)O[Cr]([O-])(=O)=O KMUONIBRACKNSN-UHFFFAOYSA-N 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 210000000476 body water Anatomy 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 238000002835 absorbance Methods 0.000 description 2
- 238000007796 conventional method Methods 0.000 description 2
- 238000003912 environmental pollution Methods 0.000 description 2
- 238000011017 operating method Methods 0.000 description 2
- 239000012286 potassium permanganate Substances 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- VEXZGXHMUGYJMC-UHFFFAOYSA-M Chloride anion Chemical compound [Cl-] VEXZGXHMUGYJMC-UHFFFAOYSA-M 0.000 description 1
- 241001269238 Data Species 0.000 description 1
- 238000000862 absorption spectrum Methods 0.000 description 1
- 230000002378 acidificating effect Effects 0.000 description 1
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 238000005443 coulometric titration Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 239000013505 freshwater Substances 0.000 description 1
- 150000002500 ions Chemical class 0.000 description 1
- 238000005461 lubrication Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000013011 mating Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- QJGQUHMNIGDVPM-UHFFFAOYSA-N nitrogen group Chemical group [N] QJGQUHMNIGDVPM-UHFFFAOYSA-N 0.000 description 1
- 239000001301 oxygen Substances 0.000 description 1
- 229910052760 oxygen Inorganic materials 0.000 description 1
- 238000005375 photometry Methods 0.000 description 1
- 238000003825 pressing Methods 0.000 description 1
- 239000013535 sea water Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000010561 standard procedure Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/33—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using ultraviolet light
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/18—Water
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/18—Water
- G01N33/1806—Water biological or chemical oxygen demand (BOD or COD)
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24137—Distances to cluster centroïds
- G06F18/2414—Smoothing the distance, e.g. radial basis function networks [RBFN]
Abstract
The invention discloses automatic water quality monitoring system and monitoring methods under a kind of multifunctional water based on improved RBFNN algorithm, system includes host computer and monitor, host computer includes Labview interface part and Python data processing section, and monitor includes scalable fixed bracket, box body, slave computer, sliding block and power module；When monitor works, current water sample is titrated with different reagents, and control ultraviolet source and issue different light to monitor different water quality parameters；Slave computer controls sliding block simultaneously, moves up and down box body in vertical direction, and then acquires multiple water quality parameters of different water depth；Host computer completes Data Analysis Services and display.Sample data is divided into test set and training set by the present invention, with training set training improved RBFNN model, is carried out offline classification processing to test set sample data with trained improved RBFNN model, is helped preferably to analyze water quality situation.
Description
Technical field
The present invention relates to water monitoring devices and method, more particularly, to a kind of based on the multi-functional of improved RBFNN algorithm
Underwater automatic water quality monitoring system and monitoring method.
Background technique
With the development of industry, water pollution is increasingly severe, and COD (COD) and nitrogen content, pH value can be anti-
The degree polluted in water by reducing substances is reflected, is the major criterion for measuring environmental quality, currently, the measurement conventional method of water quality
Generally use chemical method, mainly use potassium dichromate standard method, coulometric titration, according to water quality chlorine ion concentration it is different its
Monitoring method is also different, and fresh water generally uses potassium dichromate method, acidic potassium permanganate method, and seawater uses basic potassium permanganate method.
But these conventional methods all have that reagent dosage is big, and secondary pollution is serious, the testing time is long, operating procedure is many and diverse, measurement object
Disadvantage single, the degree of automation is low has been not suitable with the market demand of modernization.
The spectrophotometry de termination of water quality measurement is carried out in the Fundamentals of Measurement of titration, and absorption spectrum original is used
Reason, according to after the completion of being titrated in solution in reagent ion concentration how much, the ultraviolet source measurement that different wave length is respectively adopted is different
The absorbance of the water quality of concentration solution to be measured, therefore according to gained absorbance working curve, can represent chemistry in water sample needs
Oxygen amount, nitrogen content, pH value.Spectrophotometry de termination is a kind of improvement of titration, saves reagent, easy to operate.But it is existing
Instrument can only measure a kind of parameter of water quality mostly, and cannot complete the unified measurement of multi-parameter, this also results in water quality measurement
As a result one-sidedness and inaccuracy, therefore how to study a kind of water quality monitoring instrument of novel support multi parameter simultaneous measuring
It is extremely urgent.
The existing water quality monitor based on spectrophotometry in market generally requires the dedicated examination of mating corresponding production company
Agent, and artificial titer reagent is had to, manual sampling is needed, manual operation completes each water quality monitoring step, therefore wastes
A large amount of manpower, material resources and financial resources, and the water sample obtained is single, cannot embody the water quality situation under the water environment of different depth,
Monitoring result is undesirable, and compatibility of the water quality monitoring instrument under different measurement environment is very poor, for water environment
Actual measurement is unilateral, and measuring speed and precision are to be improved, and instrument price is expensive, bulky operating procedure complexity is not easy to
Practical application.Therefore how to realize collect multi-functional, cheap, data dynamic access and processed offline, in-site measurement at a distance
Monitoring, manual measurement and automatic measurement, different water depth water quality measurement, measurement are simple accurately and fast portable in being integrally water quality
One of the pressing issues that monitoring instrument is faced.
Summary of the invention
Goal of the invention: a purpose is to provide automatic water quality monitoring under a kind of multifunctional water based on improved RBFNN algorithm
System and monitoring method extend water quality monitor function, improve water quality to solve the prior art and apply the existing above problem
Monitoring instrument measures multiple spot water quality to the automatic measurement function and field measurement function of the multiple parameters under different water depth
The many kinds of parameters of water quality at different water depth, so that water quality monitoring work is more convenient, meanwhile, it is designed and is used using Labview
Family gui interface, and Python is called to realize the classification scheduling algorithm to water quality measurement data, it completes accurate to the multiple spot of water quality parameter
Acquisition and data analysis.
Technical solution: for achieving the above object, the invention adopts the following technical scheme:
Automatic water quality monitoring system under a kind of multifunctional water based on improved RBFNN algorithm, including host computer and monitor,
Host computer includes Labview interface part and Python data processing section, and host computer designs GUI circle by Labview
Face, by internal interface function call Python function using improved RBFNN algorithm completed sample according to processing, first by sample number
It is divided into test set and training set according to collection, improved RBFNN model is trained with training set, with trained improved RBFNN model to survey
Examination collection sample data carries out offline classification processing, and the sample data is to read host computer from SD card data dynamic access module
, by multiple data sets for measuring obtained water quality parameter and forming；And it designs gui interface and result is shown；
Monitor includes scalable fixed bracket, box body, slave computer, sliding block and power module, in which:
Scalable fixed pedestal lower end is fixed in water, for fixing entire monitor；
Box body sealing, top are provided with a hole, and box portion outside is equipped with a measurement switch, and inboard wall of cartridge is equipped with waterproof layer, and
Slave computer and power module are located inside waterproof layer；
The water quality parameter data of acquisition are uploaded to by slave computer for controlling sliding block and acquisition water quality parameter data
Host computer；
Sliding block is for connecting box body and scalable fixed bracket；
Power module is each module for power supply of monitor；
Slave computer controls sliding block, moves up and down box body in vertical direction, and then acquires the multiple of different water depth
Water quality parameter.
Optionally, scalable fixed bracket includes upper and lower two sections stainless steel iron pipe, and upper section is telescopic joint, and lower section is to fix
Section, telescopic joint lower end have a night bolt, multiple fixation holes with night bolt cooperation are distributed on fixed knot, when bullet
When spring bolt is inserted into different fixation holes, entire scalable fixed bracket has different height, and telescopic joint and fixed knot pass through
Night bolt is fixedly connected with fixation hole.
Optionally, slave computer includes single-chip microcontroller, multichannel spectrophotometry water quality module and data access module, monolithic
Machine includes control unit, outside RTC, external crystal-controlled oscillation, signal condition amplifying circuit and multichannel ADC conversion module, multichannel point
Light photometric measurement water quality module include firm banking, binary channels spectrophotometric device, test tube fixing device, silicon photocell sensor 1,
Silicon photocell sensor 2, ultraviolet source 1, ultraviolet source 2, titration slot and water valve, binary channels spectrophotometric device are fixed on fixed bottom
On seat, it is equipped with optical path on-off switch in test tube fixing device inner wall, the test tube for holding test agent is placed on test tube and fixes
In device, and test tube is open face box body top drilling, and water valve is set to test tube fixing device upper end, when water quality to be measured fill to
When liquid level is more than water valve, control unit controls water valve closure, opens titration slot and titrates current invisible spectro water quality；External RTC
It is connect with control unit with external crystal-controlled oscillation, control unit controls two ultraviolet sources by output PWM wave and issues varying strength
And the ultraviolet light of frequency, ultraviolet light pass through spectrophotometric device inner passage respectively and are irradiated in corresponding silicon photocell sensor,
Then optical signal is converted to electric signal by silicon photocell sensor, and input signal conditioning amplifying circuit carries out the electric signal
Enhanced processing finally passes through multichannel ADC conversion module input control cell processing.
In another embodiment of the present invention, automatic water quality monitoring system under a kind of multifunctional water based on improved RBFNN algorithm
Monitoring method, comprising the following steps:
(1) monitor is first debugged before starting measurement: scalable fixed stent length being adjusted to adapt to the height of current level
Scalable fixed bracket is inserted under water by degree, and it is firm fixed to insert；
(2) judge whether measurement switch is opened, if measurement switch has been opened, to control unit input the current depth of water and
Box body declines unit height value every time, and executes step (3)；If it is not, then continuing to execute step (2)；
(3) before starting monitoring, judge whether present cassette height is more than or equal to the water depth value of input, if before starting monitoring
It finds that current falling head is more than or equal to the water depth value of input, then illustrates that step (2) input is wrong, need return step
(2) the current depth of water and box body are re-entered and declines unit height value every time；If it is not, thening follow the steps (4)；
(4) control unit control stepper motor rotates forward, and so that box body is declined a unit height value, and monitor current depth of water position
The different quality parameter value set；
(5) one-shot measurement terminates, and control unit recalculates and judges currently whether falling head has been more than or equal to input
Water depth value, if not, then it represents that instrument reaches the bottom not yet, and return step (4) executes；If so, thening follow the steps (6)；
(6) control unit control stepper motor reversion, makes box body rise a unit height value；
(7) control unit calculates and judges currently to have gone up the water depth value whether height is more than or equal to input, if it is not, then table
Show that instrument reaches the water surface or more not yet, return step (6) executes；If so, indicate that box body has arrived at the water surface or more, this
Measurement terminates；
(8) after to be measured, SD card data dynamic access module is connected on host computer, epigynous computer section is from SD card
Data dynamic access module reads measurement data, and carries out sample data processing by internal interface function call Python, first
Sample data set is divided into test set and training set, improved RBFNN model is trained with training set, with trained improved RBFNN
Model carries out offline classification processing to test set sample data, and designs gui interface and show to result.
Further, the method for multiple water quality parameter values of current depth of water position is monitored in step (4) are as follows:
(41) after box body reaches commitment positions, water valve is opened in control unit control, allows extraneous water quality above box body
Water inlet flow into test tube in, if the priming charge matter water surface reach water valve more than, control unit control water valve close, reach water intaking
The purpose of sample；
(42) control unit is current according to currently needing the different reagent of the water quality parameter measured control titration slot to titrate
Water sample closes titration slot after the completion of titration；
(43) control unit output frequency be i PWM wave to ultraviolet source, make its sending corresponding frequencies and intensity it is ultraviolet
Light measures current water quality parameter α using spectrophotometry, surveys n times respectively with two silicon photocell sensors, remove maximum value and
Its average value is taken after minimum value, and current water quality parameter result is respectively stored in SD card；Wherein water quality parameter α is water quality
COD concentration, nitrogen content or pH value, i indicate to get the frequency of PWM wave required for current water quality parameter α.
Further, in the step (8) with training set training improved RBFNN model, and in advance trained improvement
The method that RBFNN model carries out offline classification processing to test set sample data includes:
(81) cluster centre c is adjusted with least square error criterionj, traditional two step method RBFNN is optimized；
(82) U (c is uniformly adjusted with gradient descent methodj,σj,ωij) value, the RBFNN finally trained, to traditional two steps
Method RBFNN carries out double optimization, c in formulajIndicate cluster centre, σjIndicate diameter sound stage width degree, ωijOutput layer weight is indicated, by three
One vector set U (c of a parameterj,σj,ωij) indicate；
(83) by existing water quality sample data set standardize, establish index system, i.e., by water quality sample data according to
The difference of measurement object is divided into: the inhomogeneities such as measurement COD, measurement nitrogen content, measurement pH value；
(84) test that the sample data set read out from SD card dynamic access module is divided into 1/3 by reserving method
Collect the training set with 2/3, and adds water pollution grade label manually to data in training set according to the parameter value measured；
(85) improved RBFNN is trained with training set, and is tested on test set, test result can join according to inside water quality
Number characteristic is divided into multiple water pollution grades automatically, and the results are shown on gui interface.
Further, cluster centre c is adjusted with least square error criterion in above-mentioned steps (81)j, to traditional two step method
The process that RBFNN is optimized is,
(810) firstly, determining the initial configuration at hidden layer center using the clustering algorithm based on dynamic attenuation radius, while benefit
With sample information dynamic control min cluster radius；
(811) then, it is based on error sum of squares criterion, it is influenced by investigating sample movement, adjusts center point value,
Diameter sound stage width degree is determined using class spacing combination sample actual distribution in class simultaneously；
(812) finally, determining the weight of hidden layer and output layer using pseudoinverse technique, ω can finally be obtainedij；
(813) after calculating by the optimization of first time, the basic network topology and overall network for having obtained RBFNN are joined
Number, network parameter include cluster centre, diameter sound stage width degree and output layer weight, by three parameters with a vector set U (cj,σj,
ωij) indicate, then pass through the parameter of the unified training adjustment RBFNN model of gradient descent method.
Further, in above-mentioned steps (810) dynamic attenuation radius clustering algorithm are as follows:
(a) initialization sample data set, i.e. training set or test set；
(b) each sample point that selection sample data is concentrated, calculates separately at a distance from existing cluster centre, finds it most
The cluster centre of neighbour；
(c) cluster radius is calculated, judges whether to meet the Gauss distance between sample point and the cluster centre and is less than cluster
Radius, if it is less, the sample point is added in this cluster；
(d) sample point is calculated every time it is necessary to subtract this sample point in original data set, until original
Sample point in data set all calculates one time, and otherwise terminator goes to second step, continue to calculate next sample point.
Further, the method for the fine tuning at hidden layer center is carried out in above-mentioned steps (811) using error sum of squares criterion
Are as follows:
(a) dynamic attenuation cluster result is utilized, initial error quadratic sum is calculatedFormula
In, xtIndicate sample data, cjIndicate current cluster centre；
(b) it to each sample in each cluster, calculates Enable pil=minj≠i{pij, piiIndicate spacing in class, pijIndicate this
Spacing between sample point and other cluster centres；
If (c) pil< pij, then sample is moved into l class, modifies cluster centre, and calculate new error sum of squares:
Jc+1=Jc-(pii-pil)；
If (d) Jc+1< Jc, then (2) step is returned to, otherwise, algorithm terminates.
Further, specific calculating step of the gradient descent method in this innovatory algorithm in above-mentioned steps (813) are as follows:
(a) mean square deviation between training result and legitimate reading is calculatedM is in formula
Input sample number, L are output node sum,For training result,For legitimate reading, RBFNN training goal is exactly as far as possible
The value of ground reduction E；
(b) when the value of E is also not up to training requirement index and the frequency of training also not up to setting upper limit, following formula is calculated:η is the step-length in training gradient descent method, and η is bigger, and gradient declines faster, training result
Easily convergence, but η value should be reduced when error change is little, prevent over-fitting and reduce fluctuating error, therefore preferably with
Newton method adds momentum method to improve algorithm, so that η value is variable,For the gradient descent direction of algorithm, determined by E, U(τ)With U(τ+1)The respectively τ and U (c of τ+1 time calculatingj,σj,ωij) value.
The utility model has the advantages that compared with prior art, present invention employs at the spectrophotometry of multi-path measurement different water depth
COD, nitrogen content, pH value, realize automation equipment, avoid secondary pollution, save the cost, improve water quality monitoring accuracy and complete
While the property of face, a large amount of manpower, material resources and financial resources are saved, meanwhile, instrument supports in-site measurement, long-range monitoring, is particularly suitable for open country
Outer water quality measurement；Instrument is effectively reduced irrelevant factor and water quality is surveyed by using multi-channel measurement different quality parameter
The interference of amount makes to measure more accurate quick；SD card data dynamic access function is introduced, makes to operate simpler convenience.
Existing water quality monitoring instrument has only completed single data acquisition function, for the later period data processing not
Too many is related to, and since the parameter of water quality monitoring is numerous, it is difficult to add suitable label to sample data to assess water quality
Quality, the present invention in improved RBFNN algorithm be suitable for such data set classification and data processing work, by training set
The study of sample and data inwardness and rule are explained to the classification of test set sample, both can be used for finding in data
Distributed architecture, also can be used as classification etc. other learning tasks forerunner.Equipment instrument is small simultaneously is convenient for carrying, and price is low
It is honest and clean, it is more close to practical application, industry and demands of individuals is met to a certain extent, helps further to improve environmental pollution
Problem has very high social utility value and the value of environmental protection.
Detailed description of the invention
Fig. 1 is monitor structural schematic diagram of the present invention；
Fig. 2 is monitor circuit connection diagram of the present invention；
Fig. 3 is measurement flow chart of the invention；
Fig. 4 is the method flow diagram of measurement different quality parameter of the invention；
Fig. 5 is water quality sample data sorting algorithm flow chart of the invention；
Fig. 6 is the algorithm flow chart of improved RBFNN of the invention.
Specific embodiment
Technical solution of the present invention is described in detail in the following with reference to the drawings and specific embodiments.
The present invention implements automatic water quality monitoring system and monitoring side under a kind of multifunctional water based on improved RBFNN algorithm
Method can be realized water-quality COD concentration, nitrogen content, pH value at multichannel spectrophotometry measurement different water depth, automatically complete not
With at the depth of water sampling of water quality, titration, parameter measurement, SD card data dynamic access, at in-site measurement and off-line monitoring and data
Manage function.Wherein STM32F103RBT6 is master chip (i.e. control unit, similarly hereinafter), coordinates each module and orderly works.
Automatic water quality monitoring system under a kind of multifunctional water based on improved RBFNN algorithm, including host computer and monitor,
Host computer includes Labview interface part and Python data processing section, and host computer designs GUI circle by Labview
Face, by internal interface function call Python function using improved RBFNN algorithm completed sample according to processing.The host computer
Python data processing section opens Python by Labview internal interface function Open Python Session, and leads to
It crosses Create Session function coding and completes subsequent data processing operation, pass through Python Node function call Python foot
This, terminates Python with Close Python Session function and calls, prevent memory overflow.Using Create Session
Function can encode realization related algorithm after opening the python program board program editing page wherein.The sample data
It is the data sets for reading host computer from data access module, being made of the water quality parameter that multiple measurements obtain, and with new
Water quality parameter increase, this data set is dynamically changeable.
Monitor includes scalable fixed bracket, box body, slave computer, sliding block and power module.Apparatus measures are automatic
Change, measurement parameter is comprehensive, and measurement result is simply accurate, is particularly suitable for water quality at offline water quality monitoring and field different water depth and supervises
It surveys.Wherein, scalable fixed pedestal lower end is fixed in water, for fixing entire monitor；Box body sealing, top is provided with one
Hole, box portion outside are equipped with a measurement switch, and inboard wall of cartridge is equipped with waterproof layer, and slave computer and power module are located in waterproof layer
Face；Slave computer is built in box body, for controlling sliding block and acquisition water quality parameter data, and by the water quality parameter number of acquisition
According to being uploaded to host computer；Sliding block is for connecting box body and scalable fixed bracket；Power module is built in box body, including
DC power supply drive module and voltage amplification module, the output of DC power supply drive module are connected with the input of voltage amplification module, electricity
Pressure amplification module output end is connect with each modular power source input terminal respectively, is each module for power supply of monitor.
As depicted in figs. 1 and 2, water quality automonitor under the multifunctional water based on improved RBFNN algorithm, wherein can stretch
The fixed bracket that contracts is made of the height-adjustable stainless steel iron pipe of two sections, its length can be adjusted according to the depth of water, described scalable solid
The adjustable length long enough of fixed rack and guarantee highest point are more than the distance of at least one box body height of the water surface, to guarantee box body
It is on the water surface in initial position；Scalable fixed bracket is divided into two sections up and down, and upper section is telescopic joint, and lower section is to fix
Section, telescopic joint lower end have a night bolt, multiple fixation holes are distributed with every same distance on fixed knot, night bolt
When being inserted into different fixation holes, entire scalable fixed bracket has different height, only needs when needing to adjust height by bullet
Spring bolt, which is pressed, then can move up and down telescopic joint in fixation hole, adjust scalable fixed support height.When box body needs to decline
When, it is only necessary to stepper motor rotation is controlled, then box body can be by fixed pulley and movable pulley along scalable fixed bracket decline one
Section distance.
The box body only has water inlet to have opening, remaining four sides seals, and prevents instrument in box body built-in water
Waterlogged damage circuit is provided with slave computer and power module in box body, and box body top is additionally provided with measurement switch, for starting and
Close monitor；Box body is connected by sliding block with adjustable stationary barrier.
The slave computer includes single-chip microcontroller, multichannel spectrophotometry water quality module and data access module, single-chip microcontroller
As the core board of instrument, it is placed in box body bottom, is connected respectively with other modules, data processing is completed and order controls, it is described
Single-chip microcontroller includes control unit, outside RTC, external crystal-controlled oscillation, signal condition amplifying circuit and multichannel ADC conversion module, described
Multichannel spectrophotometry water quality module includes firm banking, binary channels spectrophotometric device, test tube fixing device, silicon photocell
Sensor 1, silicon photocell sensor 2, ultraviolet source 1, ultraviolet source 2, titration slot and water valve, binary channels spectrophotometric device are fixed
On the fixed base, it is equipped with optical path on-off switch in test tube fixing device inner wall, the test tube for holding test agent is placed on
In test tube fixing device, and its face box body water inlet that is open, water valve are set to test tube fixing device upper end, when water quality to be measured fills
When completely to liquid level more than water valve, control unit controls water valve closure, opens the current invisible spectro water quality of titration slot titration；It is external
RTC and external crystal-controlled oscillation are connect with control unit, and control unit is issued different strong by output PWM wave two ultraviolet sources of control
It spends and the ultraviolet light of frequency, binary channels spectrophotometric device middle part is opened there are two the optical channel placed is intersected, the both ends of optical channel are divided
Not Fang Zhi ultraviolet source and silicon photocell sensor, ultraviolet light pass through respectively spectrophotometric device inner passage be irradiated to it is corresponding
In silicon photocell sensor, then optical signal is converted to electric signal by silicon photocell sensor, and then input signal conditioning is put
Big circuit amplifies processing to the electric signal, finally passes through multichannel ADC conversion module input control cell processing；Using two
A ultraviolet source can measure the water quality parameter of different data range respectively, titrate in slot and be placed with different enough titer reagents,
Test agent is formed for titrating water quality stoste, the pin of water valve and control unit is connected directly and is directly controlled by control unit
It is opened and closed.Described control unit using ARM handle chip STM32F103RBT6, the signal condition amplifying circuit include by
The voltage amplifier circuit and rc filter circuit of OPA2336UA chip composition, amplification and filtering for analog signal；It is described
Multichannel ADC conversion module is equipped with AD7705 high precision analogue conversion chip, and acquisition precision is up to 16, using SPI mode
It is communicated with control unit, and uses dma mode, realize that multi channel signals acquire simultaneously.
The data access module includes SD card data dynamic access module and EEPROM data cache module, SD card data
Dynamic access module is bi-directionally connected with control unit and host computer respectively, and the EEPROM data cache module and control unit connect
It connects；Complete the offline storage of data.The power module is built in box body, including DC power supply drive module and voltage amplification
Module, DC power supply drive module output with voltage amplification module input connect, voltage amplification module output end respectively with each mould
The connection of block power input is each module for power supply of instrument.The voltage amplification module is powered by DC power supply drive module, according to
Multiple PT1301 chips separately design amplifying circuit, and respectively disparate modules provide operating voltage.The DC power supply drives mould
Block by two section 1.5v dry cell batteries at.
The sliding block includes stepper motor, fixed pulley and movable pulley, and fixed pulley is fixed on scalable support bracket fastened
Top, movable pulley are fixed on box body on the lateral wall of scalable fixed bracket side, and close with scalable fixed bracket
Fitting, stepper motor are fixed on box body top, are socketed on stepper motor output shaft by the draught line that fixed pulley is drawn, if stepping
Motor rotates clockwise or counter-clockwise, and will drive box body and moves in vertical direction, and movable pulley moves down in vertical direction
Dynamic, movable pulley plays the movement of lubrication box body and box body is fixed to scalable fixation and prop up while moving up and down with box body
Effect on frame.Wherein, stepper motor fixing end is welded on above box portion left, and box body passes through sliding block and scalable fixation
Bracket connection, by control stepper motor rotation, reach adjustment oneself height, measurement different water depth at water quality situation mesh
's.
As shown in figure 3, under a kind of multifunctional water based on improved RBFNN algorithm automatic water quality monitoring system monitoring water quality
Method it is as follows:
(1) monitor is first debugged before starting measurement: scalable fixed stent length being adjusted to adapt to the height of current level
Scalable fixed bracket is inserted under water by degree, and it is firm fixed to insert；
(2) judge whether measurement switch is opened, if measurement switch has been opened, to control unit input the current depth of water and
Box body declines unit height value every time, and executes step (3)；If it is not, then continuing to execute step (2)；
(3) before starting monitoring, whether the current falling head of judgement is more than or equal to the water depth value of input, if starting to monitor
Before find that current falling head is more than or equal to the water depth value of input, then illustrate that step (2) input is wrong, need return step
(2) the current depth of water and box body are re-entered and declines unit height value every time；If it is not, thening follow the steps (4)；
(4) control unit control stepper motor rotates forward, and so that box body is declined a unit height value, and monitor current depth of water position
The different quality parameter value set；
(5) one-shot measurement terminates, and control unit recalculates and judges currently whether falling head has been more than or equal to input
Water depth value, if not, then it represents that instrument reaches the bottom not yet, and return step (4) executes；If so, thening follow the steps (6)；
(6) control unit control stepper motor reversion, makes box body rise a unit height value；
(7) control unit calculates and judges currently to have gone up the water depth value whether height is more than or equal to input, if it is not, then table
Show that instrument reaches the water surface or more not yet, return step (6) executes；If so, indicate that box body has arrived at the water surface or more, this
Measurement terminates；
(8) after to be measured, SD card data dynamic access module is connected on host computer, epigynous computer section is from SD card
Data dynamic access module reads measurement data, and carries out sample data processing by internal interface function call Python, first
Sample data set is divided into test set and training set, improved RBFNN model is trained with training set, with trained improved RBFNN
Model carries out offline classification processing to test set sample data, and designs gui interface and show to result.By improving
Test set sample data after RBFNN algorithm process can be divided into automatically multiple water pollutions according to water quality inner parameter characteristic
Grade, user can according to be divided in the test set sample data of the different quality class of pollution to different location and different water depth at
Water quality distinguish, the water pollution situation at each place of distinguishing judgement and the depth of water, can also by water quality compared with
The sample data in poor region is analyzed, it is found that this main influence factor of region water pollution is which kind of parameter of water quality
(COD, PH, nitrogen content etc.).
Referring to fig. 4, the specific method and step of different quality parameter are measured in the step (4) are as follows:
(41) after box body reaches commitment positions, water valve is opened in control unit control, allows extraneous water quality above box body
Water inlet flow into test tube in, if the priming charge matter water surface reach water valve more than, control unit control water valve close, reach water intaking
The purpose of sample；
(42) control unit is current according to currently needing the different reagent of the water quality parameter measured control titration slot to titrate
Water sample closes titration slot after the completion of titration；
(43) control unit output frequency be i PWM wave to ultraviolet source, make its sending corresponding frequencies and intensity it is ultraviolet
Light measures current water quality parameter α using spectrophotometry, surveys n times respectively with two silicon photocell sensors, remove maximum value and
Its average value is taken after minimum value, and current water quality parameter result is respectively stored in SD card；Wherein water quality parameter α is water quality
COD concentration, nitrogen content or pH value, i indicate to get the frequency of PWM wave required for current water quality parameter α.
Referring to Fig. 5, with training set training improved RBFNN model in above-mentioned steps (8), and with trained improvement in advance
The method that RBFNN model carries out offline classification processing to test set sample data includes:
It include: to adjust cluster centre c with least square error criterionj, traditional two step method RBFNN is optimized；Use gradient
Descent method uniformly adjusts U (cj,σj,ωij) value, the RBFNN finally trained, to traditional two step method RBFNN two suboptimums of progress
Change, c in formulajIndicate cluster centre, σjIndicate diameter sound stage width degree, ωijOutput layer weight is indicated, by three parameters, one vector set
Close U (cj,σj,ωij) indicate；By standardizing to existing water quality sample data set, index system is established, i.e., by water quality sample number
It is divided into according to according to the difference of measurement object: the inhomogeneities such as measurement COD, measurement nitrogen content, measurement pH value；It will be from SD card dynamic access
The sample data set that module is read out is divided into 1/3 test set and 2/3 training set by reserving method, and according to measuring
Parameter value adds water pollution grade label to data in training set manually；Improved RBFNN is trained with training set, and is being tested
It is tested on collection, test result can be divided into automatically multiple water pollution grades according to water quality inner parameter characteristic, and will as the result is shown
On gui interface.
Concrete application are as follows:
(1) water quality sample data set is imported, because clustering method can be used for the classification problem of data untagged collection, number
It can be no label data collection according to collection.
(2) data set standardizes, and can handle data set using the methods of min-max, i.e., with some sample in data set
Data are divided by the difference of maximum number and minimum number in data set as the data after the sample standardization.
(3) use reserves method and data set is divided into 1/3 test set and 2/3 training set,
(4) using improved RBFNN training method training RBFNN model, and with trained improved RBFNN model to survey
Examination collection carries out offline classification processing.
(5) the pollution level label of different test set water quality datas is obtained.It is available than tradition with improved RBFNN algorithm
Clustering algorithm or the more accurate classification data and pollution level label of traditional two step method RBFNN, it is higher to obtain confidence level
Classification results, user can be eliminated and artificially be judged according to the pollution level of the result accurate judgement test set water quality data
Process avoids randomness, blindness, subjectivity and the inaccuracy of artificial judging result, ensure that result is reliable, accurate.
It preferably, is min-max method to the standardized standardized method of available data collection, circular isWherein, x is new data, x1For former data, min is the minimum value in data set, max be in data set most
Big value.
Referring to Fig. 6, cluster centre c is adjusted with least square error criterionj, mistake that traditional two step method RBFNN is optimized
Cheng Wei,
(1) it firstly, determining the initial configuration at hidden layer center using the clustering algorithm based on dynamic attenuation radius, utilizes simultaneously
Sample data dynamic control min cluster radius.This improvement had both prevented radii fixus to cluster poor adaptivity, avoided
Multiple explorations of decaying least radius empirical value determines, and effectively reduces the caused overfitting phenomenon of infinite radius reduction
It generates.
(2) then, it is based on error sum of squares criterion, it is influenced by investigating sample movement, adjusts center point value, together
Class spacing combination sample actual distribution determines diameter sound stage width degree in Shi Liyong class.This process has fully considered that class spacing is poly- to sample
It is excessively overlapped to avoid the specification area that unified diameter sound stage width degree may cause for the influence of class.
(3) finally, determining the weight of hidden layer and output layer using pseudoinverse technique, ω can finally be obtainedij。
(4) after calculating by the optimization of first time, the basic network topology and overall network parameter of RBFNN has been obtained
(cluster centre, diameter sound stage width degree, output layer weight), by three parameters with a vector set U (cj,σj,ωij) indicate.Then
Pass through the parameter of the unified training adjustment RBFNN model of gradient descent method.
Preferably, the clustering algorithm of the dynamic attenuation radius are as follows:
(1) initialization sample data set (training set or test set).
(2) each sample point that selection sample data is concentrated, calculates separately at a distance from existing cluster centre, finds it most
The cluster centre of neighbour.
(3) cluster radius is calculated, judges whether to meet the Gauss distance between sample point and the cluster centre and is less than cluster
Radius, if it is less, the sample point is added in this cluster.
(4) sample point is calculated every time it is necessary to subtract this sample point in original data set, until original
Sample point in data set all calculates one time, and otherwise terminator goes to second step, continue to calculate next sample point.
Preferably, the method for the fine tuning that hidden layer center is carried out using error sum of squares are as follows:
(1) dynamic attenuation cluster result is utilized, initial error quadratic sum is calculatedFormula
In, xtIndicate sample data, cjIndicate current cluster centre.
(2) it to each sample in each cluster, calculates Enable pil=minj≠i{pij, piiIndicate spacing in class, pijIndicate this
Spacing between sample point and other cluster centres.
(3) if pil< pij, then sample is moved into l class, modifies cluster centre, and calculate new error sum of squares:
Jc+1=Jc-(pii-pil)。
(4) if Jc+1< Jc, then (2) step is returned to, otherwise, algorithm terminates.
Preferably, specific calculating step of the gradient descent method in this improved RBFNN algorithm are as follows:
(1) mean square deviation between training result and legitimate reading is calculatedM is in formula
Input sample number, L are output node sum,For training result,For legitimate reading, RBFNN training goal is exactly as far as possible
The value of ground reduction E.
(2) when the value of E is also not up to training requirement index and the frequency of training also not up to setting upper limit, following formula is calculated:η is the step-length in training gradient descent method, and η is bigger, and gradient declines faster, training result
Easily convergence, but η value should be reduced when error change is little, prevent over-fitting and reduce fluctuating error, therefore preferably with
Newton method adds momentum method to improve algorithm, so that η value is variable,For the gradient descent direction of algorithm, determined by E, U(τ)With U(τ+1)The respectively τ and U (c of τ+1 time calculatingj,σj,ωij) value.
In short, the present invention realizes the COD, nitrogenous at the spectrophotometry measurement different water depth of multi-path using STM32
Amount, pH value realize automation equipment, avoid secondary pollution, save the cost, improve water quality monitoring accuracy and comprehensive same
When, a large amount of manpower, material resources and financial resources are saved, meanwhile, instrument supports in-site measurement, long-range monitoring, is particularly suitable for field water survey
Amount；Instrument is effectively reduced irrelevant factor for the interference of water quality measurement by using multi-channel measurement different quality parameter,
Make to measure more accurate quick；SD card data dynamic access function is introduced, makes to operate simpler convenience；Using improvement
RBFNN algorithm carries out offline classification processing to test data, ensure that the reliability and accuracy of data classification result, so that this
Monitoring analysis system is more close to practical application；Equipment instrument is small simultaneously is convenient for carrying, and cheap, is more close to reality
Border application, meets industry and demands of individuals to a certain extent, helps further to improve problem of environmental pollution, has very high
Social utility's value and the value of environmental protection.
Claims (10)
1. automatic water quality monitoring system under a kind of multifunctional water based on improved RBFNN algorithm, it is characterised in that: including host computer
And monitor, host computer include that Labview interface part and Python data processing section, host computer pass through Labview
Gui interface is designed, by internal interface function call Python function using improved RBFNN algorithm completed sample according to processing,
Sample data set is first divided into test set and training set, improved RBFNN model is trained with training set, with trained improvement
RBFNN model carries out offline classification processing to test set sample data, and the sample data is from SD card data dynamic access mould
Data sets that block reads host computer, being made of the water quality parameter that multiple measurements obtain；And it designs gui interface and result is carried out
Display；
Monitor includes scalable fixed bracket, box body, slave computer, sliding block and power module, in which:
Scalable fixed pedestal lower end is fixed in water, for fixing entire monitor；
Box body sealing, top are provided with a hole, and box portion outside is equipped with a measurement switch, and inboard wall of cartridge is equipped with waterproof layer, and the next
Machine and power module are located inside waterproof layer；
The water quality parameter data of acquisition are uploaded to upper by slave computer for controlling sliding block and acquisition water quality parameter data
Machine；
Sliding block is for connecting box body and scalable fixed bracket；
Power module is each module for power supply of monitor；
Slave computer controls sliding block, moves up and down box body in vertical direction, and then acquires multiple water quality of different water depth
Parameter.
2. automatic water quality monitoring system under a kind of multifunctional water based on improved RBFNN algorithm according to claim 1,
Be characterized in that: scalable fixed bracket includes upper and lower two sections stainless steel iron pipe, and upper section is telescopic joint, and lower section is fixed knot, is stretched
It saves lower end and has a night bolt, multiple fixation holes with night bolt cooperation are distributed on fixed knot, when night bolt
When being inserted into different fixation holes, there is entire scalable fixed bracket different height, telescopic joint and fixed knot to be inserted by spring
Pin is fixedly connected with fixation hole.
3. automatic water quality monitoring system under a kind of multifunctional water based on improved RBFNN algorithm according to claim 1,
Be characterized in that: slave computer includes that single-chip microcontroller, multichannel spectrophotometry water quality module and data access module, single-chip microcontroller include
Control unit, outside RTC, external crystal-controlled oscillation, signal condition amplifying circuit and multichannel ADC conversion module, multichannel spectrophotometric
Measuring water quality module includes firm banking, binary channels spectrophotometric device, test tube fixing device, silicon photocell sensor 1, silicon photoelectricity
Pond sensor 2, ultraviolet source 1, ultraviolet source 2, titration slot and water valve, binary channels spectrophotometric device is fixed on the fixed base,
It is equipped with optical path on-off switch in test tube fixing device inner wall, the test tube for holding test agent is placed on test tube fixing device
It is interior, and test tube opening face box body top drilling, water valve is set to test tube fixing device upper end, when water quality to be measured is filled to liquid level
When more than water valve, control unit controls water valve closure, opens titration slot and titrates current invisible spectro water quality；External RTC and outer
Portion's crystal oscillator is connect with control unit, and control unit controls two ultraviolet sources by output PWM wave and issues varying strength and frequency
The ultraviolet light of rate, ultraviolet light pass through spectrophotometric device inner passage respectively and are irradiated in corresponding silicon photocell sensor, then
Optical signal is converted to electric signal by silicon photocell sensor, and input signal conditioning amplifying circuit amplifies the electric signal
Multichannel ADC conversion module input control cell processing is finally passed through in processing.
4. automatic water quality monitoring system under a kind of described in any item multifunctional waters based on improved RBFNN algorithm of claim 1-3
The monitoring method of system, which comprises the following steps:
(1) monitor is first debugged before starting measurement: scalable fixed stent length being adjusted to adapt to the height of current level, it will
Scalable fixed bracket insertion is underwater, and it is firm fixed to insert；
(2) judge whether measurement switch is opened, if measurement switch has been opened, input the current depth of water and box body to control unit
Decline unit height value every time, and execute step (3)；If it is not, then continuing to execute step (2)；
(3) before starting monitoring, judge whether present cassette height is more than or equal to the water depth value of input, if just sending out before starting monitoring
Now current falling head is more than or equal to the water depth value of input, then illustrates that step (2) input is wrong, need return step (2) weight
It newly inputs the current depth of water and box body and declines unit height value every time；If it is not, thening follow the steps (4)；
(4) control unit control stepper motor rotates forward, and so that box body is declined a unit height value, and monitor current depth of water position
Different quality parameter value；
(5) one-shot measurement terminates, and control unit recalculates and judges whether current falling head is more than or equal to the water of input
Deep value, if not, then it represents that instrument reaches the bottom not yet, and return step (4) executes；If so, thening follow the steps (6)；
(6) control unit control stepper motor reversion, makes box body rise a unit height value；
(7) control unit calculates and judges currently to have gone up the water depth value whether height is more than or equal to input, if not, then it represents that instrument
Device reaches the water surface or more not yet, and return step (6) executes；If so, indicate that box body has arrived at the water surface or more, this measurement
Terminate；
(8) after to be measured, SD card data dynamic access module is connected on host computer, epigynous computer section is from SD card data
Dynamic access module reads measurement data, and carries out sample data processing by internal interface function call Python, first by sample
Notebook data collection is divided into test set and training set, improved RBFNN model is trained with training set, with trained improved RBFNN model
Offline classification processing is carried out to test set sample data, and designs gui interface and result is shown.
5. water quality automatic monitoring method under a kind of multifunctional water based on improved RBFNN algorithm according to claim 4,
It is characterized in that, the method for multiple water quality parameter values of the current depth of water position of monitoring in step (4) are as follows:
(41) when box body reach commitment positions after, control unit control open water valve, allow extraneous water quality above box body into
The mouth of a river flows into test tube, if the priming charge matter water surface reaches water valve or more, control unit controls water valve and closes, and reaches water sampling
Purpose；
(42) control unit is according to currently needing the water quality parameter measured the control titration slot current water sample of different reagent titration,
Titration slot is closed after the completion of titration；
(43) PWM wave that control unit output frequency is i makes it issue the ultraviolet light of corresponding frequencies and intensity to ultraviolet source,
Current water quality parameter α is measured using spectrophotometry, surveys n times respectively with two silicon photocell sensors, removes maximum value and most
Its average value is taken after small value, and current water quality parameter result is respectively stored in SD card；Wherein water quality parameter α is water quality
COD concentration, nitrogen content or pH value, i indicate to get the frequency of PWM wave required for current water quality parameter α.
6. the monitoring of a kind of multiple spot Monitoring And Analysis of The Quality system based on improved RBFNN algorithm according to claim 4 point
Analysis method, which is characterized in that in the step (8) with training set training improved RBFNN model, and in advance trained improvement
The method that RBFNN model carries out offline classification processing to test set sample data includes:
(81) cluster centre c is adjusted with least square error criterionj, traditional two step method RBFNN is optimized；
(82) U (c is uniformly adjusted with gradient descent methodj, σj, ωij) value, the RBFNN finally trained, to traditional two step method
RBFNN carries out double optimization, c in formulajIndicate cluster centre, σjIndicate diameter sound stage width degree, ωijOutput layer weight is indicated, by three
One vector set U (c of parameterj, σj, ωij) indicate；
(83) by standardizing to existing water quality sample data set, index system is established, i.e., by water quality sample data according to measurement
The difference of object is divided into: the inhomogeneities such as measurement COD, measurement nitrogen content, measurement pH value；
(84) by the sample data set read out from SD card dynamic access module by reserve method be divided into 1/3 test set with
2/3 training set, and water pollution grade label is added manually to data in training set according to the parameter value measured；
(85) improved RBFNN is trained with training set, and is tested on test set, test result can be special according to water quality inner parameter
Property is divided into multiple water pollution grades automatically, and the results are shown on gui interface.
7. the monitoring of a kind of multiple spot Monitoring And Analysis of The Quality system based on improved RBFNN algorithm according to claim 6 point
Analysis method, it is characterised in that: adjust cluster centre c with least square error criterioni, traditional two step method RBFNN is optimized
Process is,
(1) firstly, determining the initial configuration at hidden layer center using the clustering algorithm based on dynamic attenuation radius, while sample is utilized
Information dynamic control min cluster radius；
(2) then, it is based on error sum of squares criterion, it is influenced by investigating sample movement, adjusts center point value, while benefit
Diameter sound stage width degree is determined with class spacing combination sample actual distribution in class；
(3) finally, determining the weight of hidden layer and output layer using pseudoinverse technique, ω can finally be obtainedij；
(4) after calculating by the optimization of first time, the basic network topology and overall network parameter of RBFNN, network have been obtained
Parameter includes cluster centre, diameter sound stage width degree and output layer weight, by three parameters with a vector set U (cj, σj, ωij) table
Show, then passes through the parameter of the unified training adjustment RBFNN model of gradient descent method.
8. the monitoring of a kind of multiple spot Monitoring And Analysis of The Quality system based on improved RBFNN algorithm according to claim 7 point
Analysis method, it is characterised in that: the clustering algorithm of the dynamic attenuation radius are as follows:
(1) initialization sample data set, i.e. training set or test set；
(2) each sample point that selection sample data is concentrated, calculates separately at a distance from existing cluster centre, finds its arest neighbors
Cluster centre；
(3) cluster radius is calculated, judges whether to meet the Gauss distance between sample point and the cluster centre and is less than cluster radius,
If it is less, the sample point is added in this cluster；
(4) sample point is calculated every time it is necessary to subtract this sample point in original data set, until original data
The sample point of concentration all calculates one time, and otherwise terminator goes to second step, continue to calculate next sample point.
9. the monitoring of a kind of multiple spot Monitoring And Analysis of The Quality system based on improved RBFNN algorithm according to claim 7 point
Analysis method, it is characterised in that: the method for the fine tuning that hidden layer center is carried out using error sum of squares criterion are as follows:
(1) dynamic attenuation cluster result is utilized, initial error quadratic sum is calculatedIn formula, xt
Indicate sample data, cjIndicate current cluster centre；
(2) it to each sample in each cluster, calculates Enable pil=minj≠i{pij, piiIndicate spacing in class, pijIndicate this
Spacing between sample point and other cluster centres；
(3) if pil< pij, then sample is moved into l class, modifies cluster centre, and calculate new error sum of squares: Jc+1=
Jc-(pii-pil)；
(4) if Jc+1< Jc, then (2) step is returned to, otherwise, algorithm terminates.
10. the monitoring of a kind of multiple spot Monitoring And Analysis of The Quality system based on improved RBFNN algorithm according to claim 7 point
Analysis method, it is characterised in that: specific calculating step of the gradient descent method in this innovatory algorithm are as follows:
(1) mean square deviation between training result and legitimate reading is calculatedM is input in formula
Sample number, L are output node sum,For training result,For legitimate reading, RBFNN training goal is exactly to subtract as much as possible
The value of small E；
(2) when the value of E is also not up to training requirement index and the frequency of training also not up to setting upper limit, following formula is calculated:η is the step-length in training gradient descent method, and η is bigger, and gradient declines faster, training result
Easily convergence, but η value should be reduced when error change is little, prevent over-fitting and reduce fluctuating error, therefore preferably with
Newton method adds momentum method to improve algorithm, so that η value is variable,For the gradient descent direction of algorithm, determined by E, U(τ)With U(τ+1)The respectively τ and U (c of τ+1 time calculatingj, σj, ωij) value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811138411.1A CN109270017A (en) | 2018-09-28 | 2018-09-28 | Automatic water quality monitoring system and monitoring method under a kind of multifunctional water based on improved RBFNN algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811138411.1A CN109270017A (en) | 2018-09-28 | 2018-09-28 | Automatic water quality monitoring system and monitoring method under a kind of multifunctional water based on improved RBFNN algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109270017A true CN109270017A (en) | 2019-01-25 |
Family
ID=65198676
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811138411.1A Pending CN109270017A (en) | 2018-09-28 | 2018-09-28 | Automatic water quality monitoring system and monitoring method under a kind of multifunctional water based on improved RBFNN algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109270017A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2002074694A2 (en) * | 2001-03-16 | 2002-09-26 | Ewatertek Inc. | System and method for monitoring water quality and transmitting water quality data |
| CN103175513A (en) * | 2013-03-01 | 2013-06-26 | 戴会超 | System and method for monitoring hydrology and water quality of river basin under influence of water projects based on Internet of Things |
| CN103399130A (en) * | 2013-07-30 | 2013-11-20 | 哈尔滨工业大学 | Portable tap-water quality monitoring device and monitoring method thereof |
| CN106290770A (en) * | 2016-09-14 | 2017-01-04 | 中国农业大学 | The chpn monitoring method of a kind of water quality and system |
| CN107024366A (en) * | 2016-12-22 | 2017-08-08 | 中国科学院遥感与数字地球研究所 | A kind of Portable water sampler |
| CN207295272U (en) * | 2017-08-08 | 2018-05-01 | 惠州市德康兴家居用品有限公司 | A kind of flexible clothes hanger |
| CN108549000A (en) * | 2018-03-27 | 2018-09-18 | 华南理工大学 | A kind of on-line monitoring equipment of breeding water body health water quality |
-
2018
- 2018-09-28 CN CN201811138411.1A patent/CN109270017A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2002074694A2 (en) * | 2001-03-16 | 2002-09-26 | Ewatertek Inc. | System and method for monitoring water quality and transmitting water quality data |
| CN103175513A (en) * | 2013-03-01 | 2013-06-26 | 戴会超 | System and method for monitoring hydrology and water quality of river basin under influence of water projects based on Internet of Things |
| CN103399130A (en) * | 2013-07-30 | 2013-11-20 | 哈尔滨工业大学 | Portable tap-water quality monitoring device and monitoring method thereof |
| CN106290770A (en) * | 2016-09-14 | 2017-01-04 | 中国农业大学 | The chpn monitoring method of a kind of water quality and system |
| CN107024366A (en) * | 2016-12-22 | 2017-08-08 | 中国科学院遥感与数字地球研究所 | A kind of Portable water sampler |
| CN207295272U (en) * | 2017-08-08 | 2018-05-01 | 惠州市德康兴家居用品有限公司 | A kind of flexible clothes hanger |
| CN108549000A (en) * | 2018-03-27 | 2018-09-18 | 华南理工大学 | A kind of on-line monitoring equipment of breeding water body health water quality |
Non-Patent Citations (3)
| Title |
|---|
| 田津 等: "基于三阶段RBFNN 学习算法的复杂样本分类研究", 《系统工程与电子技术》 * |
| 肖清湄: "基于LM改进的RBF神经网络算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 陈金龙: "基于GPRS的COD值远程监控系统研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105388309B (en) | The automatic quick determination method of trace iron ions and system and application in Power Plant Water Vapor | |
| CN101183071B (en) | Novel water quality analysis meter | |
| CN205426779U (en) | On --spot autoanalyzer of sulphion | |
| CN203365316U (en) | Multi-parameter water quality analyzer | |
| CN206248652U (en) | Real-time in-situ water quality monitor | |
| CN108760642A (en) | The real-time Water Test Kits of full spectrum | |
| CN109270016A (en) | Automatic water quality monitoring system and monitoring method under a kind of multifunctional water based on clustering algorithm | |
| CN101556245B (en) | Chlorophyll measurement method based on RGB digital signal | |
| CN205080143U (en) | Automatic quick detecting system of trace iron ion in power plant's steam | |
| CN112730299B (en) | Gas-oil ratio measuring method and device based on underground infrared spectroscopy | |
| CN205449792U (en) | Water sampling device and monitoring system thereof | |
| CN214011063U (en) | Nutritive salt analyzer | |
| CN204479564U (en) | Based on the semisubmersible Water Test Kits of wireless telecommunications | |
| CN109270017A (en) | Automatic water quality monitoring system and monitoring method under a kind of multifunctional water based on improved RBFNN algorithm | |
| CN110124761A (en) | Water environment multi-parameter electrochemical detection device and its detection method based on micro-fluidic chip | |
| CN107364550B (en) | Online automatic detection ship for fishery water quality | |
| CN206460030U (en) | The integrated automonitor of city black and odorous water | |
| CN105651739B (en) | Contents of many kinds of heavy metal ion nanocomposite optical detection device and method based on Stripping Voltammetry | |
| CN202024965U (en) | Real-time on-line detecting device for concentration of nitrate ions in seawater | |
| CN217605650U (en) | Continuous flow analysis system for chemical oxygen demand in water | |
| CN109270015A (en) | A kind of multiple spot Monitoring And Analysis of The Quality system and method for monitoring and analyzing based on improved RBFNN algorithm | |
| CN109270014A (en) | Water quality automonitor and monitoring method under a kind of multifunctional water | |
| CN214150541U (en) | Soil nitrate nitrogen real-time detection system based on electrochemistry | |
| CN109060692A (en) | Active phosphorus automatic analyzer and its measuring method based on syringe pump | |
| CN205404410U (en) | Double -light -path method littoral zone water chlorophyll normal position monitoring devices |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190125 |
|
| RJ01 | Rejection of invention patent application after publication |