CN109165247A - Sewage measurement data intelligence preprocess method - Google Patents
Sewage measurement data intelligence preprocess method Download PDFInfo
- Publication number
- CN109165247A CN109165247A CN201811162519.4A CN201811162519A CN109165247A CN 109165247 A CN109165247 A CN 109165247A CN 201811162519 A CN201811162519 A CN 201811162519A CN 109165247 A CN109165247 A CN 109165247A
- Authority
- CN
- China
- Prior art keywords
- data
- feature vector
- input feature
- sewage
- amount
- 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.)
- Granted
Links
Abstract
A kind of sewage measurement data intelligence preprocess method of disclosure of the invention.The method includes the following steps: 1) to pre-process the data predicted or controlled for sewage quality based on time series；2) model of mind is determined, target characteristic amount in input feature vector amount and preprocessed data in preprocessed data；3) input feature vector amount in preprocessed data is grouped according to the position of actual spot of measurement；4) every group of input feature vector amount is traversed using previous group input feature vector amount as object of reference, and traversal unit is a sampling period, and exports the data after combination；5) model of mind training is carried out using the data of predetermined ratio after combination；6) square mean error amount (MSE) of model of mind is calculated using data remaining after combination；7) time interval of every group of input feature vector amount is updated；8) if every group of time interval is less than or equal to the time of entire sewage process, return step 4)；9) the smallest time interval of square mean error amount is exported in ergodic process to get to the time difference between each input feature vector amount；10) data are reconfigured according to the time interval between each input feature vector amount, for model of mind training.
Description
Technical field
The invention belongs to survey when sewage data processing field more particularly to a kind of modeling using artificial intelligence technology to sewage
It measures data and carries out pretreated method.
Background technique
In China, most cities sewage treatment plant all uses activated sludge process to handle sewage.Activated sludge process
Sewage disposal process be a close coupling multiple-input and multiple-output dynamical system, have large time delay, non-linear, time-varying and
The features such as uncertain, brings difficulty to modeling.
Traditionally, people can use acquired abundant data knowledge and model to real system, and in each control
It is used widely in field processed.However, for the continuous of complexity, time-varying, large time delay and non-linear effluent control system, mathematics
The method of modeling is by stern challenge.In recent years, with the development of artificial intelligence, it can use its technology and number measured to sewage
According to black box mechanism model is established, this class model does not need to have a clear understanding of microbial reaction process mechanism, it is only necessary to research process
In output and input data, choose effective model, the time series mould that system both can be obtained be adjusted to model parameter
Type.However, and sewage disposal system itself has time lag, causes since the measurement of sewage data is based on synchronization
The sequence problem that sewage measurement feature is analyzed during intelligent modeling becomes a difficult point.
Summary of the invention
To overcome drawbacks described above, the purpose of the present invention is to provide a kind of sewage measurement data intelligence preprocess methods.This
Method can intelligently export the time interval between each input feature vector amount and target characteristic amount, solve sewage feature it is big when
Stagnant problem.
In order to achieve the above objectives, the present invention provides a kind of method of sewage measurement data Intelligent treatment, comprising: Xia Shubu
It is rapid:
1) data predicted or controlled for sewage quality based on time series are pre-processed；
2) model of mind is determined, target characteristic amount in input feature vector amount and preprocessed data in preprocessed data；
3) input feature vector amount in preprocessed data is grouped according to the position of actual spot of measurement；
4) every group of input feature vector amount is traversed using previous group input feature vector amount as object of reference, and traversal unit is one and adopts
The sample period, and export the data after combination；
5) model of mind training is carried out using the data of predetermined ratio after combination；
6) square mean error amount (MSE) of model of mind is calculated using data remaining after combination；
The definition of square mean error amount:
Wherein,----square mean error amount；nsamples----total sample number；
yi----ith measurement value；----i-th predicted value；
7) time interval of every group of input feature vector amount is updated；
8) if every group of time interval is less than or equal to the time of entire sewage process, return step 4)；
9) export the smallest time interval of square mean error amount in ergodic process to get between each input feature vector amount when
Between it is poor；
10) data are reconfigured according to the time interval between each input feature vector amount, for model of mind training.
Wherein, the step 1) specifically:
Obtain the acquisition data predicted or controlled for sewage quality based on time series；
Missing Data Filling, 3delta processing and difference processing are carried out to data
The present invention can intelligently calculate the time interval between sewage measurement data, provide for the training of model of mind
The preprocessed data of high quality.A kind of new approaches are provided for the data processing based on time series, simplify model of mind
Complexity.
Detailed description of the invention
Fig. 1 is the flow diagram of intelligent data processing method of the present invention.
Specific implementation method
Technical solution of the present invention is further illustrated below with reference to the actual operating data of certain sewage treatment plant.
S101, the acquisition data predicted or controlled for sewage quality based on time series are obtained；
Specifically, acquisition data are the detection data in sewage disposal process, this detection data is every two hours to adopt automatically
Collect the data that report, mainly include flow of inlet water, influent COD, influent ammonia nitrogen, dissolved oxygen, sludge activity, water temperature, water flow,
It is discharged COD and water outlet ammonia nitrogen.Note that all data processings below are all based on time progress, so every acquisition and recording
In all should include data acquisition time.
S102, data are carried out with Missing Data Filling, 3delta processing and difference processing；
Specifically, currently the detection data of sewage treatment plant's missing is carried out at missing values supplement using Lagrangian method
Reason is replaced the data beyond 3delta range using 3delta boundary value, and finally to treated, data carry out difference
Processing, i.e., subtract last moment with the data of subsequent time.
S103, model of mind, input feature vector amount and target characteristic amount are determined；
Specifically, water outlet ammonia nitrogen is predicted using BP neural network model, i.e., target signature is water outlet ammonia nitrogen, input
Feature has flow of inlet water, influent COD, influent ammonia nitrogen, dissolved oxygen, sludge activity, water temperature.
S104, input feature vector amount is grouped according to the position of actual spot of measurement；
Specifically, the actual spot of measurement acquired according to sewage treatment plant's data, is divided into 3 groups for upper face data, first group is
Flow of inlet water, influent COD and influent ammonia nitrogen, second group is dissolved oxygen, sludge activity and water temperature, and third group is water outlet ammonia nitrogen.
S105, every group of input feature vector amount are traversed using previous group input feature vector amount as object of reference, and traversal unit is one
A sampling period, and export the data after combination；
S106, model of mind training is carried out using data a certain proportion of after combination (such as 80% data)；
S107, the square mean error amount (MSE) that model of mind is calculated using remaining data (such as 20% data)；
The definition of square mean error amount:
Wherein,----square mean error amount；nsamples----total sample number
yi----ith measurement value----i-th predicted value
S108, the time interval for updating every group of input feature vector amount；
If S109, every group of time interval are less than or equal to the time of entire sewage process, return step S105；
Specifically, it is assumed that the entire sewage process period of current sewage treatment plant is 24 hours, then every group when
Between be spaced and be less than equal to 24 hours.Using first group of data (flow of inlet water, influent COD, influent ammonia nitrogen) as object of reference, to
Two groups of data (dissolved oxygen, sludge activity, water temperature) are traversed, and traversal range is 0~12, similarly third group data (water outlet ammonia
Nitrogen) it is traversed using second group of data as object of reference, traversal range is also 0~12.Data after each traversal assembling are pressed
Collect { Data_Test } as verifying according to rear 20% data, other data are split as training set { Data_Train }.So
Parameter adjustment is carried out to BP model using training set { Data_Train } afterwards, utilizes verifying collection { Data_Test } after the completion of model
It is predicted, and exports the square mean error amount of prediction, until having traversed.
The smallest time interval of square mean error amount is in S110, output ergodic process to get between each input feature vector amount
Time difference；
S111, data are reconfigured according to the time interval between each input feature vector amount, for model of mind training.
Specifically, finding out the smallest time interval of square mean error amount in ergodic process is [4,2], i.e., to certain sewage treatment plant
It carries out needing to carry out offset assembly operation to data when water outlet ammonia nitrogen prediction, the data after reconfiguring is supplied to intelligent mould
Type.For example need to predict the water outlet ammonia nitrogen of 2018-06-02 12:00:00, then needing to input 2018-06-02 00:00:00
Flow of inlet water, influent COD and the influent ammonia nitrogen and dissolved oxygen, the sludge activity at 2018-06-02 08:00:00 moment at moment
And water temperature.
The above is only better embodiment of the invention, should not be construed as limiting the scope of the invention, and
And the scope of the claims that the present invention is advocated is not limited merely to this, all personages for being familiar with this field skill, according to the present invention
Disclosed technology contents, can think easily and equivalence changes, protection scope of the present invention should all be fallen into.For example, to combination
When data are divided, division proportion is not limited in 8:2；Judging quota is not limited to square mean error amount.
Claims (3)
1. a kind of sewage measurement data intelligence preprocess method, which is characterized in that the method includes the following steps:
1) data predicted or controlled for sewage quality based on time series are pre-processed；
2) model of mind is determined, target characteristic amount in input feature vector amount and preprocessed data in preprocessed data；
3) input feature vector amount in preprocessed data is grouped according to the position of actual spot of measurement；
4) every group of input feature vector amount is traversed using previous group input feature vector amount as object of reference, and traversal unit is a sampling week
Phase, and export the data after combination；
5) model of mind training is carried out using the data of predetermined ratio after combination；
6) square mean error amount (MSE) of model of mind is calculated using data remaining after combination；
The definition of square mean error amount:
Wherein,----square mean error amount；nsamples----total sample number；yi----ith measurement value；
----i-th predicted value；
7) time interval of every group of input feature vector amount is updated；
8) if every group of time interval is less than or equal to the time of entire sewage process, return step 4)；
9) the smallest time interval of square mean error amount is exported in ergodic process to get to the time between each input feature vector amount
Difference；
10) data are reconfigured according to the time interval between each input feature vector amount, for model of mind training.
2. sewage measurement data intelligence preprocess method as described in claim 1, which is characterized in that the step 1) is specific
Are as follows:
Obtain the acquisition data predicted or controlled for sewage quality based on time series；
Missing Data Filling and 3delta processing are carried out to acquisition data；
To treated, data carry out difference processing.
3. sewage measurement data intelligence preprocess method as claimed in claim 2, which is characterized in that the acquisition data are
Detection data in sewage disposal process, this detection data are data that every two hours automatic collection reports, including flow of inlet water,
Influent COD, influent ammonia nitrogen, dissolved oxygen, sludge activity, water temperature, water flow, water outlet COD and water outlet ammonia nitrogen.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811162519.4A CN109165247B (en) | 2018-09-30 | 2018-09-30 | Intelligent pretreatment method for sewage measurement data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811162519.4A CN109165247B (en) | 2018-09-30 | 2018-09-30 | Intelligent pretreatment method for sewage measurement data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109165247A true CN109165247A (en) | 2019-01-08 |
| CN109165247B CN109165247B (en) | 2021-07-23 |
Family
ID=64877250
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811162519.4A Active CN109165247B (en) | 2018-09-30 | 2018-09-30 | Intelligent pretreatment method for sewage measurement data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109165247B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101923083A (en) * | 2009-06-17 | 2010-12-22 | 复旦大学 | Sewage chemical oxygen demand soft measuring method based on support vector machine and neural network |
| CN102616927A (en) * | 2012-03-28 | 2012-08-01 | 中国科学技术大学 | Adjusting method of technological parameters of sewage treatment and device |
| CN103399134A (en) * | 2013-08-20 | 2013-11-20 | 渤海大学 | Sewage COD soft measurement method based on output observer |
| CN103942457A (en) * | 2014-05-09 | 2014-07-23 | 浙江师范大学 | Water quality parameter time series prediction method based on relevance vector machine regression |
| CN107169610A (en) * | 2017-06-02 | 2017-09-15 | 中国农业大学 | Aquaculture dissolved oxygen prediction method and device |
| JP2018018413A (en) * | 2016-07-29 | 2018-02-01 | ペンタフ株式会社 | Sewage unknown water simple evaluation method |
| CN107688871A (en) * | 2017-08-18 | 2018-02-13 | 中国农业大学 | A kind of water quality prediction method and device |
| CN108414016A (en) * | 2018-03-01 | 2018-08-17 | 深圳市晟达机械设计有限公司 | A kind of sewage network monitoring system based on big data technology |
-
2018
- 2018-09-30 CN CN201811162519.4A patent/CN109165247B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101923083A (en) * | 2009-06-17 | 2010-12-22 | 复旦大学 | Sewage chemical oxygen demand soft measuring method based on support vector machine and neural network |
| CN102616927A (en) * | 2012-03-28 | 2012-08-01 | 中国科学技术大学 | Adjusting method of technological parameters of sewage treatment and device |
| CN103399134A (en) * | 2013-08-20 | 2013-11-20 | 渤海大学 | Sewage COD soft measurement method based on output observer |
| CN103942457A (en) * | 2014-05-09 | 2014-07-23 | 浙江师范大学 | Water quality parameter time series prediction method based on relevance vector machine regression |
| JP2018018413A (en) * | 2016-07-29 | 2018-02-01 | ペンタフ株式会社 | Sewage unknown water simple evaluation method |
| CN107169610A (en) * | 2017-06-02 | 2017-09-15 | 中国农业大学 | Aquaculture dissolved oxygen prediction method and device |
| CN107688871A (en) * | 2017-08-18 | 2018-02-13 | 中国农业大学 | A kind of water quality prediction method and device |
| CN108414016A (en) * | 2018-03-01 | 2018-08-17 | 深圳市晟达机械设计有限公司 | A kind of sewage network monitoring system based on big data technology |
Non-Patent Citations (1)
| Title |
|---|
| 叶延亮 等: "污水预警中的化学需氧量(COD)预测技术", 《化工自动化及仪表》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109165247B (en) | 2021-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110054274B (en) | Water purification flocculation precipitation dosing control method | |
| CN108898215B (en) | Intelligent sludge bulking identification method based on two-type fuzzy neural network | |
| Côte et al. | Dynamic modelling of the activated sludge process: improving prediction using neural networks | |
| CN109685314B (en) | Non-intrusive load decomposition method and system based on long-term and short-term memory network | |
| CN103197539B (en) | The method of wastewater disposal intelligent optimization control aeration quantity | |
| CN111533290B (en) | Method for generating optimal operation plan library of sewage treatment process and applying complex scene | |
| CN102778843B (en) | Operation control method of high magnetic grading process | |
| WO2021007871A1 (en) | Alumina production operation optimization system and method employing cloud-side collaboration | |
| CN109116100B (en) | It is a kind of based on coding-decoding structure electric load electricity consumption decomposition method | |
| CN115456479B (en) | Intelligent agricultural greenhouse environment monitoring system based on Internet of things | |
| CN111610301A (en) | Centralized water quality monitoring device, system and method | |
| CN108445861A (en) | A kind of goat fault detection method and system based on convolutional neural networks algorithm | |
| CN109459399A (en) | A kind of spectral water quality COD, turbidity detection method | |
| CN111307519A (en) | Self-adaptive variable-frequency automatic water collection system based on hydrodynamic force change and use method | |
| CN106706491B (en) | Intelligent detection method for membrane bioreactor-MBR water permeability | |
| CN113845205B (en) | High-salinity high-nitrogen sewage intelligent integrated desalting and denitrification control system | |
| CN101786675A (en) | Device and method for separating multi-parameter wastewater sources | |
| CN109165247A (en) | Sewage measurement data intelligence preprocess method | |
| CN113274885B (en) | Membrane pollution intelligent early warning method applied to membrane sewage treatment | |
| CN109494741A (en) | A kind of selective harmonic compensation method extracted based on specific subharmonic | |
| CN116859839A (en) | Industrial control method and device based on model training | |
| CN117170221A (en) | Artificial intelligence control system for sewage treatment | |
| CN114693493B (en) | IoT-based polluted river water ecological restoration system | |
| CN116119877A (en) | Sewage automatic treatment method and system based on Internet of things technology | |
| CN105792281B (en) | A kind of method and system of intelligent fire bolt terminal data transmission |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |