CN107298485A - It is a kind of based on method of the data model to the fault detection and diagnosis of During Industrial Wastewater Treatment Process - Google Patents
It is a kind of based on method of the data model to the fault detection and diagnosis of During Industrial Wastewater Treatment Process Download PDFInfo
- Publication number
- CN107298485A CN107298485A CN201710621800.9A CN201710621800A CN107298485A CN 107298485 A CN107298485 A CN 107298485A CN 201710621800 A CN201710621800 A CN 201710621800A CN 107298485 A CN107298485 A CN 107298485A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- data
- variable
- wastewater treatment
- 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
Classifications
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F3/00—Biological treatment of water, waste water, or sewage
- C02F3/30—Aerobic and anaerobic processes
- C02F3/302—Nitrification and denitrification treatment
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/08—Chemical Oxygen Demand [COD]; Biological Oxygen Demand [BOD]
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/10—Solids, e.g. total solids [TS], total suspended solids [TSS] or volatile solids [VS]
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/40—Liquid flow rate
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2303/00—Specific treatment goals
- C02F2303/14—Maintenance of water treatment installations
Abstract
The invention discloses a kind of based on method of the data model to the fault detection and diagnosis of During Industrial Wastewater Treatment Process.Input variable is used as using inhalant region dissolved oxygen amount, aerobic zone suspended solids content, water inlet flow, delivery port flow, COD and delivery port suspended solids content;It is normalized based on input variable, Holistic modeling is carried out to training data using PCA, is that each variable chooses pivot respectively according to the load vectors of generation, forms subspace;The T per sub-spaces is calculated respectively2Statistic, and construct SVDD models;Normalized real-time monitoring data is projected into each subspace, calculating obtains corresponding T2Statistic;Counting statistics amount in the higher dimensional space that SVDD is constructed apart from the center of circle distance, so as to judge whether the monitoring data at the moment normal.
Description
Technical field
The invention belongs to industrial chemical and process control crossing domain, it is related in During Industrial Wastewater Treatment Process, passes through measurement
Data volume reflects the method for system running state in time.
Background technology
The construction of Industrial Wastewater Treatment engineering is mainly the problem of solution drainage system sewage is arbitrarily discharged, and improves and administers
Sewage is to local ecological severe contamination problem.The displacement mass of industrial system is for the living environment of people, and social holds
Supervention exhibition, suffers from vital influence.
It is food industrial wastewater for influent quality, its biodegradability is higher.Concentration of wastewater is higher, its biochemical oxygen demand (BOD)
(Biochemical oxygen demand, BOD), COD (Chemical oxygen demand, COD), nitrogen always contain
Amount (total nitrogen, TN), phosphorus total content (total phosphorous, TP) isoconcentration are the 3-4 of ordinary municipal sewage
Times.For extensive sewage, treatment plant is using the bio-aeration technique that most technique is exactly traditional activated sludge process.This work
Skill is to right using the aerobe (including bacterium and protozoa) cultivated in same artificial environment after physical treatment course
Organic contamination matter in sewage is degraded.The technique is relatively simple, and the removal efficiency to COD, BOD and SS can reach expection
It is required that.But, this method only has single biotic environment, it is impossible to play and the reinforcing biological nature of different microorganisms and excellent
Gesture, can neither improve effective processing to high-concentration sewage, effective removal that can not be to compound fertilizer, it is therefore necessary to
Expected processing target is met using biochemical two stage treatment.The approach that biologic treating technique is followed mainly has two:One is
The microbial biomass of participation effect is improved, the probability that increase organic matter is contacted with microorganism, its means realized are to improve mixed liquor
Concentration.Two be the metabolic characteristic for playing different microorganisms advantage, and there is provided matched with superior microorganism physiological property for screening strain
Biotic environment.Make each quasi-microorganism try one's best " to divide labour with individual responsibility ", play sharpest edges to realize desired processing target.
Biological phosphate-eliminating denitrification process is to take this approach.Its denitrogenation is, by extending aeration time, longer to be disappeared using generation time
Change bacterium, be nitrate by mineralized nitrogen, recycle the amphimicrobian under anoxia condition is counter to digest bacterium nitrate nitrogen is converted into gaseous state
Nitrogen is escaped, so as to reach the purpose of NH3-N in water removal.Dephosphorization is entered first with facultative under anaerobic condition and aerobic polyP bacteria
Effective release of row phosphorus, recycle through anaerobism put the flora of phosphorus under aerobic condition everybody breed and inhale the characteristic of phosphorus, make raw sewage
In phosphorus be converted into biological cell (activated sludge), separated eventually through precipitation, rich phosphorus excess sludge excluded to realize.By
In this complete dephosphorization, denitrification process respectively anaerobism, and oxygen and it is aerobic under the conditions of carry out, so commonly referred to as
A2/ O techniques (i.e. the aerobic Oxic techniques of anaerobism Anaerobic+ anoxics Anoxic+).On the basis of this technologic improvement, design whole
Body sewage disposal system, its process flow diagram is shown in Fig. 1.
A in sewerage2The detailed process of/O parts, the pond body is made up of four parts, parallel operation.Per pond in the past extremely
It is inhalant region, anaerobic zone, anoxic zone, aerobic zone successively afterwards.Quick-fried oxygen head is installed, remaining each area is respectively mounted stirs under water in aerobic zone
Device, prevents mixed liquor formation precipitation.A2The running of/O reaction tanks has vital effect for wastewater treatment process.
Such as when water intake velocity is too fast, make to react insufficient in reaction tank, or oxygen content is not in normal range (NR) in reaction tank so that nitrogen phosphorus
Removal is insufficient, causes the water quality after processing not up to standard when serious, influence ecological environment.Therefore, how to determine to produce shape at present
State, maintained equipment are safe and stable to be run with optimization, is one of Industrial Wastewater Treatment engineering major issue in the urgent need to address.
The process model building based on data model obtains a large amount of successfully applications in the industry in recent years, solves many fortune
Row system status monitoring problem.Process monitoring based on data-driven is that process data is modeled during utilizing collection system, structure
The monitoring statisticss amount of process observation is made, so that the state of monitoring system operation, finds system operation failure and diagnosed in time.
A2Biochemical reaction in/O reaction tanks also generates mass data, and these data contain the feature letter of sewage disposal process
Breath, therefore can be produced based on During Industrial Wastewater Treatment Process and analyze data, set up procedure correlation model.Present invention employs base
Pivot analysis supporting vector test in data domain (the Loading-based Principal component of pivot are chosen in load
selection for Principal component anaylsis integrated with Support Vector
Data Description, PPCA-SVDD) modeling method, built using the process data being collected into During Industrial Wastewater Treatment Process
Vertical mathematical modeling.Choose certain sensitive pivot for each variable and constitute and become vector subspace, the change to each variable is more sensitive,
So as to possess more preferable procedure fault information extraction ability.Constructed by SVDD statistics so that process monitoring statistic has
The ability of course of reaction state, so as to reach the purpose of procedure fault detection.After failure is detected, according to process data, do
Go out corresponding variable contribution plot so that procedure fault reason is diagnosed, so as to further amendment failure so that process is extensive
Multiple normal operation.
The content of the invention
The present invention seeks to the process data using Industrial Wastewater Treatment factory there is provided the statistic of a condition monitoring come
Running is monitored, the system failure is found in time and is diagnosed.The observational variable of selection has:Inhalant region dissolved oxygen amount DO (x1,
Mg/L), aerobic zone suspended solids content SS (x2, mg/L), water inlet flow Qin(x3,m3/ s), delivery port flow Qef(x4,m3/s)、
COD CODin(x5, mg/L) and delivery port suspended solids content SSef(x6,mg/L).It regard this 6 variables as the work
The raw monitored variable of industry Waste Water Treatment, using related measuring instrumentss, direct measurement or indirectly calculating obtain x1~x6。
Commercial plant sample data is gathered, data model is built, by calculating the process monitoring statistic obtained to wastewater treatment process
Fault detection and diagnosis is realized, with the normal operation of safeguards system process.
1. the selection of raw monitored variable
In the technological process of Industrial Wastewater Treatment, A2/ O reaction tanks are the core cells in whole system.Influence reaction tank
The key factor that oxidation reaction is normally run has:Oxygen content, actual oxygen demand, the flow of water inlet in reaction tank, the stream of water outlet
Amount, reactant concentration.
It is characteristic of the invention that:
By using the sample data of process collection, the pivot analysis supporting vector data that pivot is chosen based on load are set up
Domain describes (PPCA-SVDD) model, and construction process monitoring statisticss amount, realization is monitored in real time to whole service process, so that
Procedure fault is monitored in time and is diagnosed.
Therefore, based on above analytic explanation, the monitored parameterses of the Industrial Wastewater Treatment engineering choose as follows:
(1) inhalant region dissolved oxygen amount DO (x1,mg/L)
(2) aerobic zone suspended solids content SS (x2,mg/L)
(3) water inlet flow Qin(x3,m3/s)
(4) delivery port flow Qef(x4,m3/s)
(5) COD CODin(x5,mg/L)
And delivery port suspended solids content SS (6)ef(x6,mg/L)
Six variables of the above can be directly obtained by measuring instrument.
2. the pretreatment of modeling sample
In order to eliminate the influence of dimension, pretreatment is normalized in the sample data to collection.Input variable utilizes formula (1)
It is normalized:
(1) in formula, xiIt is the actual measured value of i-th of observational variable, μiRepresent observed quantity xiCorresponding average, σiTable
Show observed quantity xiCorresponding variance,Represent the value after i-th of input variable normalization.
The representational commercial plant data of n groups are collected, wherein every group of packet [x containing input variable1,x2,x3,x4,x5,
x6], obtained after formula (1) normalizedForm modeling sample.
For the data x in t real-time collectingt, place is normalized using the mean variance obtained in formula (1) to it
Reason, it is as follows,
Wherein, μ=[μ1,μ2,…,μ6], and σ=[σ1,σ2,…,σ6]。
3. the During Industrial Wastewater Treatment Process monitoring model based on PPCA-SVDD
Assuming that the sample size of modeling sample is n, modeling data is normalized according to (1) formula first, using PPCA-SVDD
Modeling method, sets up initial model;Then, by analyzing loading matrix P and each load vectors p that PCA modelings are producediIt is right
The eigenvalue λ answeredi, the conversion weighted value calculated according to (2) formula is that each variable chooses sensitive pivot.
Wherein, pi,jFor (i, j) individual element in loading matrix.For a certain variableFor, calculate the institute corresponding to it
There is the average of weighted valueBy comparing wi,jWithChoose the load vectors corresponding to the conversion weighted value more than average, structure
Into variable xiProjector spaceMeanwhile, corresponding characteristic value also retains constitutive characteristic value matrixSo, each variable
Produce corresponding new subspace.Data are projected to each space respectively, sub- statistic can be calculated according to (3) formula
Because the subspace corresponding to each variable is more sensitive for the change of the variable, therefore work as variable
When failure changes, corresponding space faster can more accurately be detected.Normal training sample obtains corresponding to 6 changes
The statistic of amount, will according to Support Vector data description domain SVDDUtilize functionNon-linear projection is to higher dimensional space, then finds can surround data point after projection
Suprasphere as small as possible, wherein,Problem is converted to the following optimization of solution by SVDD
Problem:
Wherein, R and τ represent radius and the center of suprasphere respectively, and parameter C represents spheroid volume size and normal sample one's duty
Balance between wrong fraction, ξiCoefficient of relaxation is represented, that is, allows the misjudged probability of training sample.The antithesis of above-mentioned optimization problem
Form can be expressed as,
Wherein βiFor corresponding Lagrange multiplier, SVDD chooses 0≤βiEach sample corresponding to≤C as support to
Amount.So, the centre of sphere and radius of suprasphere can just be tried to achieve according to following formula,
Wherein,For any one in model supports vector.
There is the normalization target detection sample for including six variables for one that t is collectedIt projects to each son
Space, calculates the statistic composition statistics moment matrix of each subspaceIts square distance for arriving the centre of sphere is calculated againIt can be expressed as,
When distance of the test sample to the center of circle in feature space is more than radius obtained by training sample, it is believed that the sample
This is exceptional sample, conversely, the sample considered to be in normal condition.The design of process monitoring statistic is as follows,
When DR statistics are less than 1, it is believed that system is normal;When more than 1, that is, think that process monitoring statistic detects event
Barrier, gives a warning.Then, it should just carry out corresponding fault diagnosis and carry out the basic reason that discovery procedure is broken down, so as to repair
System.Here, using the pivot information in the subspace that can monitor failure, contribution rate is calculated, it is as follows
WhereinForIn j-th of element ForIn j-th of element,ForIn (i, j) it is individual
Element, then can calculate and obtain variable xiFor scoreContribution rate conti,j.Then, variable xiTotal contribution rate can calculate
,
Because the corresponding sensitive pivot of variable contains more fault information volumes on process, therefore based on sensitive pivot choosing
Made contribution plot is taken to provide a more accurate fault diagnosis result.
4. process monitoring statistic in line computation in industrial wastewater treatment system
The online calculation process of industrial wastewater treatment system process monitoring statistic is as shown in Figure 2.It is inhalant region dissolved oxygen amount, aerobic
Area suspended solids content SS, water inlet flow, delivery port flow, COD and delivery port suspended solids content, based on above-mentioned
Data are normalized by the direct measurement data of 6 input variables by (1) formula;Training data is carried out using PCA
Holistic modeling, is that each variable chooses pivot respectively according to the load vectors of generation, forms subspace;Calculate empty per height respectively
Between T2Statistic, and construct SVDD models;For the real time data of monitoring, the mean variance for utilizing (2) formula to produce is returned
One change is handled;Normalized real-time monitoring data is projected into each subspace, calculating obtains corresponding T2Statistic;Counting statistics
The distance apart from the center of circle in the higher dimensional space that SVDD is constructed is measured, so as to judge whether the monitoring data at the moment is normal.In addition,
When faulty in discovery system, the contribution rate of variable is calculated with formula (11) (12), fault diagnosis is realized.
Brief description of the drawings
Fig. 1 technique for treating industrial wastewater FB(flow block)s.
The online calculation process of Fig. 2 During Industrial Wastewater Treatment Process monitoring statisticss amounts.
Embodiment
The invention will be further described by the following examples:
For six variables, inhalant region dissolved oxygen amount DO (x1, mg/L), aerobic zone suspended solids content SS (x2, mg/L), water inlet
Mouth flow Qin(x3,m3/ s), delivery port flow Qef(x4,m3/ s), COD CODin(x5, mg/L) and delivery port suspension
Body content SSef(x6, mg/L), the history normal data for gathering the system of one group of 100 sample is used as training data, and one group
The real-time process being made up of 100 samples observes data.
1. pre-process sample
First group of data being made up of 200 samples to above-mentioned collection are normalized, and utilize (1) formula:x1's
Average is 3.78, and variance is 1.37;x2Average be 4.00, variance is 1.27;x3Average be 3417.48, variance is
1290.31;x4Average be 4090.31, variance is 1274.90;x5Average be 24.82, variance is 3.58;x6Average be
19.58, variance is 33.53.Calculating is normalized:
To the data x of t real-time collectingt(8) formula of utilization is normalized
Wherein, mean μ=[3.78,4.00,3417.48,4090.31,24.82,19.58], variances sigma=[1.37,
1.27,1290.31,1274.90,3.58,33.53].Here there is the form of decimal, 2 significant digits, four houses five are retained without exception
Enter to calculate.
2. the construction of statistic in the Monitoring Model based on PPCA-SVDD
Using PPCA-SVDD modeling methods, sample number is modeled for 100 training data, sample number is 100 number
According to being tested.Its specific model parameter is as follows:
(1) PCA modelings are carried out to training data first, produces six pivots, it is specific such as to calculate weighting matrix by formula (3)
Under:
According to weighted value be more than the variable to the averages of all weighted values, that is, choose the vectorial structure corresponding to the weighted value
This principle of subspace is made, the pivot result that each variable is chosen is as follows:
Variable x1:p2,p3,p5
Variable x2:p2,p3
Variable x3:p1,p6
Variable x4:p1,p6
Variable x5:p1,p2,p3,p5
Variable x6:p1,p4
Thus, 6 variables of correspondence generate 6 sub-spaces, and decentralized supervisory control is realized to each variable.Conversion per sub-spaces
Matrix is:
Space 1:
Space 2:
Space 3:
Space 4:
Space 5:
Space 6:
(2) in kernel function K (), s2=100;
(3) in SVDD models, parameter C=1.2;
(4) the SVDD models being made up of each subspace statistic, the parameter value of generation is as follows:
β value is 0.3006,0.4931,0.6545,0.8119,0.9374,0;Suprasphere radius R=0.216.
Described above by example, based on industrial wastewater treatment system, by measuring obtained inhalant region dissolved oxygen amount, good
Oxygen area suspended solids content, water inlet flow, delivery port flow, COD and delivery port suspended solids content, in real time,
Line monitoring system running status.
The model obtained by above-mentioned example, here is the test data sample that one group of real-time collecting is arrived:
Obtained through normalizing calculating:
By (10) formula, it is 0.3322 to calculate and obtain the statistic of the real-time sample.The value is less than 1, it is believed that be the moment
System is in normal condition.DR normalized set results corresponding to 100 test samples are as follows:
Overstriking data represent that the point exceedes control limit 1 in result of calculation, are abnormal data.Statistic at 81st point is
15.5309, hence it is evident that more than control limit.The process data at the moment is used to do fault diagnosis, calculates the contribution rate of 6 variables
Respectively:
0.315,2.031,0.929,2.743,46.401,0.972
From above diagnostic result, variable x5Contribution rate to this failure is maximum, therefore is the change in wastewater treatment process
Learn oxygen demand and there occurs variable.
Claims (4)
1. it is a kind of based on fault detection method of the data model to During Industrial Wastewater Treatment Process, with inhalant region dissolved oxygen amount DO (x1,mg/
L), aerobic zone suspended solids content SS (x2, mg/L), water inlet flow Qin(x3,m3/ s), delivery port flow Qef(x4,m3/ s), change
Learn oxygen demand CODin(x5, mg/L) and delivery port suspended solids content SSef(x6, mg/L) and it is used as input variable;Become based on input
Amount is normalized, and Holistic modeling is carried out to training data using PCA, is each variable point according to the load vectors of generation
Pivot is not chosen, forms subspace;The T per sub-spaces is calculated respectively2Statistic, and construct SVDD models;Will be normalized
Real-time monitoring data projects to each subspace, and calculating obtains corresponding T2Statistic;The higher-dimension that Counting statistics amount is constructed in SVDD
Apart from the distance in the center of circle in space, so as to judge whether the monitoring data at the moment is normal.
2. it is a kind of based on method for diagnosing faults of the data model to During Industrial Wastewater Treatment Process, it is the PPCA- based on claim 1
SVDD During Industrial Wastewater Treatment Process monitoring model, when DR statistics are less than 1, it is believed that system is normal;When more than 1, that is, recognize
Failure is detected for process monitoring statistic, is given a warning.Then, it should just carry out corresponding fault diagnosis and carry out discovery procedure
The basic reason of failure, so that repair system;Using the pivot information in the subspace that can monitor failure, contribution rate is calculated,
It is as follows:
<mrow>
<msub>
<mi>cont</mi>
<mrow>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<msubsup>
<mi>t</mi>
<mi>j</mi>
<mo>*</mo>
</msubsup>
<msubsup>
<mi>&lambda;</mi>
<mi>j</mi>
<mo>*</mo>
</msubsup>
</mfrac>
<msubsup>
<mi>p</mi>
<mrow>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
<mo>*</mo>
</msubsup>
<msub>
<mover>
<mi>x</mi>
<mo>&OverBar;</mo>
</mover>
<mi>i</mi>
</msub>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>11</mn>
<mo>)</mo>
</mrow>
</mrow>
WhereinForIn j-th of element ForIn j-th of element,ForIn (i, j) individual element,
It can then calculate and obtain variable xiFor scoreContribution rate conti,j;Then, variable xiTotal contribution rate can calculate,
<mrow>
<msub>
<mi>CONT</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<msubsup>
<mi>&Sigma;</mi>
<mi>j</mi>
<mi>m</mi>
</msubsup>
<msub>
<mi>cont</mi>
<mrow>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>12</mn>
<mo>)</mo>
</mrow>
</mrow>
The contribution rate of variable is calculated with formula (11), (12), so as to realize fault diagnosis.
3. it is according to claim 1 based on fault detection method of the data model to During Industrial Wastewater Treatment Process, its feature
It is the inhalant region dissolved oxygen amount, aerobic zone suspended solids content, water inlet flow, delivery port flow, COD and to go out
Mouth of a river suspended solids content, is directly measured by measuring instrumentss.
4. it is according to claim 1 based on fault detection method of the data model to During Industrial Wastewater Treatment Process, its feature
It is that the modeling data is normalized using following formula:
<mrow>
<msub>
<mover>
<mi>x</mi>
<mo>&OverBar;</mo>
</mover>
<mi>i</mi>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>&mu;</mi>
<mi>i</mi>
</msub>
</mrow>
<msub>
<mi>&sigma;</mi>
<mi>i</mi>
</msub>
</mfrac>
<mo>,</mo>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<mn>6</mn>
</mrow>
Wherein,--- the value after i-th of input variable normalization,
μi--- observed quantity xiCorresponding average,
σi--- observed quantity xiCorresponding variance;
Real-time monitoring data xt, it is normalized using following formula:
<mrow>
<msub>
<mover>
<mi>x</mi>
<mo>&OverBar;</mo>
</mover>
<mi>t</mi>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mi>x</mi>
<mi>t</mi>
</msub>
<mo>-</mo>
<mi>&mu;</mi>
</mrow>
<mi>&sigma;</mi>
</mfrac>
</mrow>
Wherein, μ --- include each element average, i.e. μ=[μ1,μ2,…,μ6];
σ --- include each element variance, i.e. σ=[σ1,σ2,…,σ6];
Gather commercial plant data, after normalization, using proposed based on the pivot analysis that load chooses pivot support to
Measure test in data domain algorithm (PPCA-SVDD), building process monitoring mathematical modeling.Using loading matrix characteristic, become for each
Amount selects pivot respectively, constitutes monitoring subspace, the normal condition T of generation2Statistic is used to build SVDD models.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710621800.9A CN107298485A (en) | 2017-07-27 | 2017-07-27 | It is a kind of based on method of the data model to the fault detection and diagnosis of During Industrial Wastewater Treatment Process |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710621800.9A CN107298485A (en) | 2017-07-27 | 2017-07-27 | It is a kind of based on method of the data model to the fault detection and diagnosis of During Industrial Wastewater Treatment Process |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107298485A true CN107298485A (en) | 2017-10-27 |
Family
ID=60133623
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710621800.9A Pending CN107298485A (en) | 2017-07-27 | 2017-07-27 | It is a kind of based on method of the data model to the fault detection and diagnosis of During Industrial Wastewater Treatment Process |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107298485A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110033105A (en) * | 2019-04-18 | 2019-07-19 | 中国人民解放军国防科技大学 | Suspension system fault detection method for unbalanced data set under complex working condition |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101403923A (en) * | 2008-10-31 | 2009-04-08 | 浙江大学 | Course monitoring method based on non-gauss component extraction and support vector description |
CN102487293A (en) * | 2010-12-06 | 2012-06-06 | 中国人民解放军理工大学 | Satellite communication network abnormity detection method based on network control |
CN103472820A (en) * | 2013-09-18 | 2013-12-25 | 哈尔滨工业大学 | Method for diagnosing propulsion system failure based on partial least squares algorithm |
CN103645249A (en) * | 2013-11-27 | 2014-03-19 | 国网黑龙江省电力有限公司 | Online fault detection method for reduced set-based downsampling unbalance SVM (Support Vector Machine) transformer |
CN205528209U (en) * | 2016-01-26 | 2016-08-31 | 临沂高新区鲁润净水设备有限公司 | Ozone purifier intelligence control system based on fuzzy neural network |
-
2017
- 2017-07-27 CN CN201710621800.9A patent/CN107298485A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101403923A (en) * | 2008-10-31 | 2009-04-08 | 浙江大学 | Course monitoring method based on non-gauss component extraction and support vector description |
CN102487293A (en) * | 2010-12-06 | 2012-06-06 | 中国人民解放军理工大学 | Satellite communication network abnormity detection method based on network control |
CN103472820A (en) * | 2013-09-18 | 2013-12-25 | 哈尔滨工业大学 | Method for diagnosing propulsion system failure based on partial least squares algorithm |
CN103645249A (en) * | 2013-11-27 | 2014-03-19 | 国网黑龙江省电力有限公司 | Online fault detection method for reduced set-based downsampling unbalance SVM (Support Vector Machine) transformer |
CN205528209U (en) * | 2016-01-26 | 2016-08-31 | 临沂高新区鲁润净水设备有限公司 | Ozone purifier intelligence control system based on fuzzy neural network |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110033105A (en) * | 2019-04-18 | 2019-07-19 | 中国人民解放军国防科技大学 | Suspension system fault detection method for unbalanced data set under complex working condition |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105510546B (en) | A kind of biochemical oxygen demand (BOD) BOD intelligent detecting methods based on self-organizing Recurrent RBF Neural Networks | |
Farhi et al. | Prediction of wastewater treatment quality using LSTM neural network | |
CN110088619A (en) | The intelligence system and method for process and assets Gernral Check-up, abnormality detection and control for waste water treatment plant or drinking water plant | |
CN104182794B (en) | Method for soft measurement of effluent total phosphorus in sewage disposal process based on neural network | |
CN107688701B (en) | WASP model-based water quality soft measurement and water eutrophication evaluation method | |
CN102854296A (en) | Sewage-disposal soft measurement method on basis of integrated neural network | |
CN106600509B (en) | Method for analyzing and judging water exchange and pollution discharge behaviors of enterprise based on basic data | |
CN110969346A (en) | Drainage basin water ecological function regional treatment demand evaluation method based on index screening | |
CN110232256A (en) | A kind of sewage disposal process monitoring method based on KPLS and RWFCM | |
CN102262147A (en) | Soft measurement method and system for effluent chemical oxygen demand (COD) of waste water treatment system | |
CN103606031A (en) | Evaluation method of kitchen waste feed product safety | |
US11365140B2 (en) | Decision support system and method for water treatment | |
CN108088974A (en) | A kind of flexible measurement method of anaerobism while denitrification methane phase process water outlet nitrate nitrogen | |
CN112417765B (en) | Sewage treatment process fault detection method based on improved teacher-student network model | |
CN110110466A (en) | A kind of sewage treatment target component prediction technique of machine learning and process simulation | |
CN114297954A (en) | Big data intelligent analysis digital management and control platform based on sewage treatment plant | |
CN203772781U (en) | Characteristic variable-based sewage total phosphorus measuring device | |
Xu et al. | Prediction and evaluation of indirect carbon emission from electrical consumption in multiple full-scale wastewater treatment plants via automated machine learning-based analysis | |
CN108549740A (en) | A kind of anaerobic system water outlet ammonia nitrogen flexible measurement method based on integrated intelligent algorithm | |
CN110232062A (en) | A kind of sewage disposal process monitoring method based on KPLS and FCM | |
CN107298485A (en) | It is a kind of based on method of the data model to the fault detection and diagnosis of During Industrial Wastewater Treatment Process | |
CN107665288A (en) | A kind of water quality hard measurement Forecasting Methodology of COD | |
CN106769748B (en) | Intelligent detection system for water permeability of membrane bioreactor-MBR (Membrane bioreactor) | |
Beraud et al. | Multiobjective genetic algorithms for the optimisation of wastewater treatment processes | |
CN109273058A (en) | A kind of composite algorism for the exceeded early warning of anaerobic processes volatile fatty acid |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20171027 |