CN115062272A - Water quality monitoring data abnormity identification and early warning method - Google Patents

Water quality monitoring data abnormity identification and early warning method Download PDF

Info

Publication number
CN115062272A
CN115062272A CN202210797902.7A CN202210797902A CN115062272A CN 115062272 A CN115062272 A CN 115062272A CN 202210797902 A CN202210797902 A CN 202210797902A CN 115062272 A CN115062272 A CN 115062272A
Authority
CN
China
Prior art keywords
water quality
data
model
monitoring data
quality monitoring
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
Application number
CN202210797902.7A
Other languages
Chinese (zh)
Inventor
严求真
叶旺
王越胜
何中杰
杨启尧
郭栋
张运涛
王军
汪惜丹
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hangzhou Dianzi University
Zhejiang University of Water Resources and Electric Power
Original Assignee
Hangzhou Dianzi University
Zhejiang University of Water Resources and Electric Power
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Hangzhou Dianzi University, Zhejiang University of Water Resources and Electric Power filed Critical Hangzhou Dianzi University
Publication of CN115062272A publication Critical patent/CN115062272A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A20/00—Water conservation; Efficient water supply; Efficient water use
    • Y02A20/152—Water filtration

Abstract

The invention discloses a water quality monitoring data abnormity identification and early warning method, which is characterized in that a modification model is constructed by dividing and recombining a water quality monitoring data set, a relevant correction function is added into the modification model, and the modified modification model is further integrated with a baseline model, so that the identification and early warning of abnormal data in water quality monitoring data are realized. The method can effectively capture the accurate long-range coupling correlation between the abnormal early warning output and the water quality monitoring data input, more effectively identify and early warn the occurrence of water quality abnormal events, improve the accuracy of early warning on the water quality abnormal data, and is not limited by the range of historical monitoring data.

Description

Water quality monitoring data abnormity identification and early warning method
Technical Field
The invention belongs to the technical field of water quality data abnormity monitoring and early warning, and relates to a water quality monitoring data abnormity identification and early warning method.
Background
In the process of monitoring the water quality environment, identification and early warning are usually required to be carried out on abnormal values in water quality monitoring data acquired by a water quality sensor, and the traditional method is based on threshold value type simple identification and early warning of national water quality standards. In order to reduce the negative influence of uncertain factors on water quality abnormity identification and early warning, the internal correlation information of water quality monitoring data needs to be mined, the characteristic extraction process of the water quality abnormity monitoring data is improved, and a water quality abnormity identification and early warning method with higher precision and more stability is provided.
In order to improve the accuracy and stability of the water quality abnormity monitoring data identification and early warning result, the existing improvement method mainly comprises an autoregressive model, a machine learning method and a multi-scale cyclic neural network, but the improvement method has the limitations. The autoregressive model is not suitable for a non-stable water quality monitoring data sequence, cannot simultaneously consider the information of the long-term trend of water quality change and the volatility of fine granularity, and is easy to have amplitude difference and other problems in a scene with fuzzy periodic rule; the traditional machine learning method is difficult to obtain a predicted value beyond the historical water quality monitoring data range, and events such as outlier prediction need to be post-processed; the multi-scale recurrent neural network adopts a hierarchical structure for modeling, and is difficult to optimize parameters and depends on the accuracy of a water quality prediction model.
In conclusion, when the existing method is applied to water quality monitoring data abnormity identification and early warning, due to the respective limitations, the water quality change trend and the occurrence of abnormal events are difficult to be effectively identified, so that the accuracy rate and the stability in the application of the water quality abnormity identification and early warning are insufficient. In order to improve the accuracy and stability of water quality abnormity identification and early warning, a more effective water quality abnormity identification and early warning method is urgently needed.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a water quality monitoring data abnormity identification and early warning method, which comprises the steps of constructing a modification model by dividing and recombining a water quality monitoring data set, correcting a prediction result of the modification model, and further integrating the modified modification model and a baseline model to obtain a final water quality abnormity identification result.
A water quality monitoring data abnormity identification and early warning method specifically comprises the following steps:
step one, preprocessing water quality monitoring data
Performing characteristic extraction on the collected water quality data with time correlation, and cleaning impurities and redundancies in the original data, X i For the ith column of characteristic data obtained after cleaning, X is again selected i Carrying out zero mean value normalization processing to obtain
Figure BDA0003732813510000028
Preferably, the data feature extraction method is a time stamp process or a discrete variable process.
Step two, calculating a water quality abnormity identification result based on the baseline model
For the water quality monitoring data set after the pretreatment in the step I
Figure BDA0003732813510000029
Resampling, inputting into baseline model constructed by long-short term memory neural network LSTM, inputting hidden variable of last moment of model into full-connection layer, and outputting abnormal result p of water quality monitoring data in specified time 0 ：
p 0 ＝σ(h n )
Wherein h is n Output results representing LSTM:
Figure BDA0003732813510000021
wherein
Figure BDA0003732813510000022
Representing the water quality monitoring data set after normalization treatment in the step one,
Figure BDA0003732813510000023
is composed of
Figure BDA0003732813510000024
The feature data vector at time t, k being the vector dimension.
Step three, calculating a water quality abnormity identification result based on the modification model
Normalizing the data set after the step one
Figure BDA0003732813510000025
Dividing the data into n parts according to the time sequence, and forming a new training data set X by the first n-a parts of data * Input after resamplingIn the modification model constructed by the tree model, the result p of the abnormal water quality monitoring data within the specified time is output i The latter a data are classified into the test data set, and a < n/2.
p i ＝Tree(X * ), 1≤i≤n/2-a
Preferably, the tree model is a LightGBM, XGboost or Catboost model.
Step four, introducing a correction function
Introducing a correction function, and correcting the output water quality abnormity identification result of the modification model in the third step:
Figure BDA0003732813510000026
wherein, w i Representing the result of the correction function, alpha representing the modulation factor for preventing triggering of a floating point exception, beta representing the data set
Figure BDA00037328135100000210
The statistical duration of (c).
Preferably, the modulation factor α is 1.
Step five, correcting the water quality abnormity identification result
Integrating the prediction result of the baseline model in the step two and the water quality abnormity recognition result corrected in the step four, and taking the result as the result p that the final water quality monitoring data is abnormal within the specified time:
Figure BDA0003732813510000027
wherein gamma and 1-gamma respectively represent the importance degree of the prediction results of the baseline model and the modification model.
Preferably, the method further comprises a sixth step of evaluating the recognition result:
and inputting the test data set into an integrated model, and measuring and evaluating the accuracy and stability of the water quality abnormity identification result by respectively adopting the accuracy and the AUC. The higher the accuracy rate and AUC, the better the performance of the water quality abnormity identification and early warning method.
The accuracy rate refers to the proportion of the abnormal events of water quality indexes such as dissolved oxygen, permanganate index, ammonia nitrogen, total phosphorus, total nitrogen and the like in water quality, which are identified as abnormal events:
Precision＝TP/(TP+FP)
in the formula, TP and FP respectively represent the number of water quality index abnormal events and water quality index normal events recognized as abnormal.
Calculating AUC values of water quality abnormity identification and early warning as follows:
Figure BDA0003732813510000031
wherein n is 1 、n 0 Respectively identifying the number of water quality abnormal events and the number of normal events; rank (i) indicates the sequence number of the i-th entry identified as abnormal; the probability scores are sorted from small to large, rank (i) at the ith position.
The invention has the following beneficial effects:
the method adopts a resampling method to construct a baseline model so as to not destroy the real distribution of water quality monitoring data as much as possible, then divides a recombined data set to construct a modification model so as to effectively utilize data information discarded by resampling, then adds an optimal correction function to correct a prediction result of the modification model so as to effectively utilize the prediction result information of the modification model, and finally integrates the modified modification model with the baseline model, thereby not only reducing the influence of abnormal values and missing values on an identification model, but also effectively learning the accurate long-range coupling correlation between abnormal early warning output and water quality monitoring data input, and further improving the accuracy and stability of water quality abnormal monitoring data identification and early warning.
Drawings
FIG. 1 is a flow chart of a water quality monitoring data anomaly identification and early warning method;
FIG. 2 is a schematic structural diagram of a water quality monitoring data anomaly identification model;
FIG. 3 is a schematic diagram of water quality monitoring data set division and recombination;
FIG. 4 is a schematic diagram of a model integration method.
Detailed Description
The invention is further explained below with reference to the drawings;
as shown in fig. 1 and 2, a water quality monitoring data identification and early warning method specifically comprises the following steps:
step one, preprocessing water quality monitoring data
The acquired water quality time sequence data is subjected to feature extraction through time stamp processing and discrete variable processing, and missing values and redundancies in original data such as basin basic data, water quality monitoring data and meteorological data are cleaned, and X is i The characteristic data of the ith column obtained after cleaning. In order to eliminate dimension influence of original data and enable different data to have comparability, normalization processing is carried out on cleaned feature data to realize equal-ratio scaling, and a result is mapped to [0,1 ]]Then mapping to the distribution with mean value of 0 and standard deviation of 1, and completing zero-mean normalization:
Figure BDA0003732813510000041
Figure BDA0003732813510000042
Figure BDA0003732813510000043
Figure BDA0003732813510000044
wherein the content of the first and second substances,
Figure BDA0003732813510000045
respectively the maximum value and the minimum value of the ith column of characteristic data,
Figure BDA0003732813510000046
is the ith column of characteristic data after being normalized by a linear function,
Figure BDA0003732813510000047
is that
Figure BDA0003732813510000048
M represents the size of the characteristic data of each column, mu, sigma 2 Respectively mean and variance of the ith column of feature data,
Figure BDA0003732813510000049
is the ith column of characteristic data after zero mean normalization.
Step two, calculating a water quality abnormity identification result based on the baseline model
To water quality monitoring data set
Figure BDA00037328135100000410
After resampling, a baseline model is constructed by using a long-short term memory neural network (LSTM), wherein the cyclic function of the LSTM is as follows:
Figure BDA00037328135100000411
Figure BDA00037328135100000412
Figure BDA00037328135100000413
wherein, the first and the second end of the pipe are connected with each other,
Figure BDA00037328135100000414
representing element-by-element vector multiplication, c t Cell state, h, for RNN at time t t Expressed as the final hidden state, W c 、U c 、b c Representing trainable network parameters, the input at each time also includingOutput h of last-time hidden layer t-1 ，
Figure BDA00037328135100000415
For a new alternative value vector, tanh (-) is a tanh function, namely, each element is valued at-1 to 1, i t ，o t ，f t Respectively an input gate, an output gate and a forgetting gate, and is calculated by the following formula:
Figure BDA00037328135100000416
Figure BDA0003732813510000051
Figure BDA0003732813510000052
wherein, sigma (·) is a sigmoid function, and the output value range is between 0 and 1; w is a group of i 、U i 、b i ，W o 、U o 、b o ，W f 、U f 、b f In turn relate to i t 、o t 、f t Trainable network parameters of an equation.
Inputting the characteristic data subjected to normalization processing in the step one into a baseline model for processing, and identifying water quality abnormity into a binary task, so that a hidden variable at the last moment of the LSTM model is input into a full-connection layer with only two output nodes, and the probability p of abnormity of the water quality monitoring data within a specified time can be output 0 ：
p 0 ＝σ(h n )
Wherein h is n For LSTM processed data set input, p 0 The water quality abnormity identification result of the baseline model is obtained.
Figure BDA0003732813510000053
Wherein
Figure BDA0003732813510000054
Representing the water quality monitoring data set after normalization treatment in the step one,
Figure BDA0003732813510000055
is composed of
Figure BDA0003732813510000056
The feature data vector at time t, k being the vector dimension.
The magnitude of the baseline model output value is minimized by 0 log p 0 -(1-y 0 )log(1-p 0 ) Solved by this two-class cross entropy loss function, where y 0 Represents p 0 A corresponding real tag.
The valid information in the baseline model is passed to the output layer through the hidden layer for a number of cycles. The output values of the baseline model are then further expanded to:
Figure BDA0003732813510000057
from c k Is transmitted to c n (k<n), only a small amount of linear interaction exists in the information, so that the effective sequence history information in the baseline model can be transmitted in a long-range distance. The baseline model may capture a probability value p 0 And time series characteristics c k 、
Figure BDA0003732813510000058
The coupling relationship between the two parts is obvious.
Step three, calculating a water quality abnormity identification result based on the modification model
Constructing a modification model by using an XGboost model, and normalizing the water quality monitoring data set obtained in the first step as shown in figure 3
Figure BDA0003732813510000059
The training set (train) part is divided into 4 parts according to the time sequence, and the first 2 parts and the first 3 parts of data are respectively formed into a new training data set
Figure BDA00037328135100000510
The last 2 and 3 data are put into a verification set (test) part, and input into a modification model after resampling to obtain the water quality abnormity identification probability p 1 、p 2 ：
Figure BDA00037328135100000511
Figure BDA0003732813510000061
Step four, introducing a correction function
And based on the Bayes optimal discrimination function of the two classifications, and making full use of the water quality abnormity identification result information of the modification model to obtain an optimal expression of the correction function. And (3) introducing a correction function to correct the output water quality abnormity identification result of the modification model in the step three:
Figure BDA0003732813510000062
Figure BDA0003732813510000063
wherein, w 1 、w 2 Represents p 1 、p 2 The modified function result of (1) is set in the present embodiment to prevent triggering floating point exception, and it is also beneficial to smooth the modified modification model, where β represents 720 hours and represents the data set
Figure BDA0003732813510000065
The length of time.
Step five, correcting the water quality abnormity identification result
As shown in fig. 4, the result p of the final water quality monitoring data with abnormality in a predetermined time is obtained by integrating the prediction result of the baseline model in the second step and the water quality abnormality recognition result of the modified model corrected in the fourth step:
p＝0.5p 0 +0.5(w 1 p 1 +w 2 p 2 )
step six, model evaluation
And respectively measuring and evaluating the accuracy and stability of the water quality abnormity identification algorithm by adopting the accuracy rate and the AUC. The higher the accuracy rate and AUC, the better the performance of the water quality abnormity identification and early warning method.
The accuracy rate refers to the proportion of the abnormal events of water quality indexes such as dissolved oxygen, permanganate index, ammonia nitrogen, total phosphorus, total nitrogen and the like in water quality, which are identified as abnormal events:
Precision＝TP/(TP+FP)
in the formula, TP and FP respectively represent the number of water quality index abnormal events and water quality index normal events recognized as abnormal.
Calculating AUC values of water quality abnormity identification and early warning as follows:
Figure BDA0003732813510000064
wherein n is 1 、n 0 Respectively identifying the number of water quality abnormal events and the number of normal events; rank (i) indicates the sequence number of the i-th entry identified as abnormal; the probability scores are sorted from small to large, rank (i) at the ith position.

Claims (8)

1. A water quality monitoring data abnormity identification and early warning method is characterized in that: the method specifically comprises the following steps:
step one, preprocessing water quality monitoring data
Performing characteristic extraction on the collected water quality data with time correlation, and cleaning impurities and redundancies in the original data, X i For the ith column of characteristic data obtained after cleaning, X is again selected i Carrying out zero mean value normalization processing to obtain
Figure FDA0003732813500000011
Step two, calculating a water quality abnormity identification result based on the baseline model
For the water quality monitoring data set after the pretreatment in the step I
Figure FDA0003732813500000012
Resampling, inputting into baseline model constructed by long-short term memory neural network LSTM, inputting hidden variable of last moment of model into full-connection layer, and outputting abnormal result p of water quality monitoring data in specified time 0 ：
p 0 ＝σ(h n )
Wherein h is n Output results representing LSTM:
Figure FDA0003732813500000013
wherein
Figure FDA0003732813500000014
Representing the water quality monitoring data set after normalization treatment in the step one,
Figure FDA0003732813500000015
is composed of
Figure FDA0003732813500000016
A feature data vector at time t, k being a vector dimension;
step three, calculating a water quality abnormity identification result based on the modification model
Normalizing the data set after the step one
Figure FDA0003732813500000017
Dividing the data into n parts according to the time sequence, and forming a new training data set X by the first n-a parts of data * Re-sampling and then deliveringInputting the result p of the abnormal water quality monitoring data in the modification model constructed by the tree model within the specified time i The data of the next part a are classified into a test data set, and a is less than n/2;
p i ＝Tree(X * ), 1≤i≤n/2-α
step four, introducing a correction function
Introducing a correction function, correcting the output water quality abnormity identification result of the modification model in the step three:
Figure FDA0003732813500000018
wherein, w i Representing the result of the correction function, alpha representing the modulation factor for preventing triggering of a floating point exception, beta representing the data set
Figure FDA0003732813500000019
The statistical duration of (2);
step five, correcting the water quality abnormity identification result
Integrating the prediction result of the baseline model in the step two and the water quality abnormity identification result after correction in the step four, and taking the result as a result p that the final water quality monitoring data is abnormal within a specified time:
Figure FDA0003732813500000021
wherein gamma and 1-gamma respectively represent the importance degree of the prediction results of the baseline model and the modification model.
2. The water quality monitoring data abnormity identification and early warning method as claimed in claim 1, wherein: the data feature extraction method is time stamp processing or discrete variable processing.
3. The water quality monitoring data abnormity identification and early warning method as claimed in claim 1, wherein: the water quality data comprises basin basic data, water quality monitoring data and meteorological data.
4. The long-term sequence prediction method based on the integration of modification models as claimed in claim 1, wherein: firstly, performing linear function normalization on the cleaned feature data, then performing zero-mean normalization, and mapping the normalized feature data to the distribution with the mean value of 0 and the standard deviation of 1:
Figure FDA0003732813500000022
Figure FDA0003732813500000023
Figure FDA0003732813500000024
Figure FDA0003732813500000025
wherein the content of the first and second substances,
Figure FDA0003732813500000026
respectively the maximum value and the minimum value of the ith column of characteristic data,
Figure FDA0003732813500000027
is the ith column of characteristic data after being normalized by a linear function,
Figure FDA0003732813500000028
is that
Figure FDA0003732813500000029
M represents the size of the characteristic data of each column, mu, sigma 2 Are respectively the ith columnThe mean and variance of the feature data,
Figure FDA00037328135000000210
is the ith column of characteristic data after zero mean normalization.
5. The long-time sequence prediction method based on the integration of the modified model as set forth in claim 1, wherein: the tree model is a LightGBM model, an XGboost model or a Catboost model.
6. The long-term sequence prediction method based on the integration of modification models as claimed in claim 1, wherein: the modulation factor α in the correction model is set to 1.
7. The long-time sequence prediction method based on the integration of modification models as claimed in any one of claims 1 to 6, wherein: the method also comprises a sixth step of evaluating the recognition result:
inputting the test data set into an integrated model, and measuring and evaluating the accuracy and stability of the water quality abnormity identification result by respectively adopting the accuracy and the AUC; the higher the accuracy rate and AUC are, the better the performance of the water quality abnormity identification and early warning method is;
the accuracy rate refers to the percentage of the number of water quality abnormal events identified as abnormal:
Precision＝TP/(TP+FP)
in the formula, TP and FP respectively represent the number of water quality index abnormal events and the number of water quality index normal events which are identified as abnormal;
calculating AUC values of water quality abnormity identification and early warning as follows:
Figure FDA0003732813500000031
wherein n is 1 、n 0 Respectively identifying the number of water quality abnormal events and the number of normal events; rank (i) indicates the sequence number of the i-th entry identified as abnormal; the probability scores are sorted from small to large, rank (i) inThe ith position.
8. The long-time sequence prediction method based on the integration of the modified model as claimed in claim 7, wherein: the indexes of the water quality monitoring data with abnormality in the specified time comprise indexes of dissolved oxygen, permanganate index, ammonia nitrogen, total phosphorus and total nitrogen in water quality.
CN202210797902.7A 2021-07-21 2022-07-06 Water quality monitoring data abnormity identification and early warning method Pending CN115062272A (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN2021108260384 2021-07-21
CN202110826038.4A CN113486303A (en) 2021-07-21 2021-07-21 Long-time sequence prediction method based on modification model integration

Publications (1)

Publication Number Publication Date
CN115062272A true CN115062272A (en) 2022-09-16

Family

ID=77942059

Family Applications (2)

Application Number Title Priority Date Filing Date
CN202110826038.4A Pending CN113486303A (en) 2021-07-21 2021-07-21 Long-time sequence prediction method based on modification model integration
CN202210797902.7A Pending CN115062272A (en) 2021-07-21 2022-07-06 Water quality monitoring data abnormity identification and early warning method

Family Applications Before (1)

Application Number Title Priority Date Filing Date
CN202110826038.4A Pending CN113486303A (en) 2021-07-21 2021-07-21 Long-time sequence prediction method based on modification model integration

Country Status (1)

Country Link
CN (2) CN113486303A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116182949A (en) * 2023-02-23 2023-05-30 中国人民解放军91977部队 Marine environment water quality monitoring system and method
CN116451142A (en) * 2023-06-09 2023-07-18 山东云泷水务环境科技有限公司 Water quality sensor fault detection method based on machine learning algorithm
CN117113264A (en) * 2023-10-24 2023-11-24 上海昊沧系统控制技术有限责任公司 Method for detecting abnormality of dissolved oxygen meter of sewage plant on line in real time
CN117171604A (en) * 2023-11-03 2023-12-05 城资泰诺(山东)新材料科技有限公司 Sensor-based insulation board production line abnormality monitoring system

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114169405A (en) * 2021-11-16 2022-03-11 王树松 Object classification method, device, equipment and storage medium

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116182949A (en) * 2023-02-23 2023-05-30 中国人民解放军91977部队 Marine environment water quality monitoring system and method
CN116182949B (en) * 2023-02-23 2024-03-19 中国人民解放军91977部队 Marine environment water quality monitoring system and method
CN116451142A (en) * 2023-06-09 2023-07-18 山东云泷水务环境科技有限公司 Water quality sensor fault detection method based on machine learning algorithm
CN117113264A (en) * 2023-10-24 2023-11-24 上海昊沧系统控制技术有限责任公司 Method for detecting abnormality of dissolved oxygen meter of sewage plant on line in real time
CN117113264B (en) * 2023-10-24 2024-02-09 上海昊沧系统控制技术有限责任公司 Method for detecting abnormality of dissolved oxygen meter of sewage plant on line in real time
CN117171604A (en) * 2023-11-03 2023-12-05 城资泰诺(山东)新材料科技有限公司 Sensor-based insulation board production line abnormality monitoring system
CN117171604B (en) * 2023-11-03 2024-01-19 城资泰诺(山东)新材料科技有限公司 Sensor-based insulation board production line abnormality monitoring system

Also Published As

Publication number Publication date
CN113486303A (en) 2021-10-08

Similar Documents

Publication Publication Date Title
CN115062272A (en) Water quality monitoring data abnormity identification and early warning method
CN115018021B (en) Machine room abnormity detection method and device based on graph structure and abnormity attention mechanism
CN112131212A (en) Hybrid cloud scene-oriented time sequence data anomaly prediction method based on ensemble learning technology
CN114386521A (en) Method, system, device and storage medium for detecting abnormality of time-series data
Yang et al. Autoregressive coefficient-invariant control chart pattern recognition in autocorrelated manufacturing processes using neural network ensemble
Lee et al. Studies on the GAN-based anomaly detection methods for the time series data
CN113486578A (en) Method for predicting residual life of equipment in industrial process
CN116738868B (en) Rolling bearing residual life prediction method
CN114385614A (en) Water quality early warning method based on Informmer model
CN110956309A (en) Flow activity prediction method based on CRF and LSTM
CN114297918A (en) Aero-engine residual life prediction method based on full-attention depth network and dynamic ensemble learning
CN111125186B (en) Data processing method and system based on questionnaire
Lei et al. Fault diagnosis of rotating machinery based on a new hybrid clustering algorithm
CN112947649B (en) Multivariate process monitoring method based on mutual information matrix projection
CN114443338A (en) Sparse negative sample-oriented anomaly detection method, model construction method and device
CN115290326A (en) Rolling bearing fault intelligent diagnosis method
CN116364203A (en) Water quality prediction method, system and device based on deep learning
CN115062686A (en) Multi-KPI (Key performance indicator) time sequence abnormity detection method and system based on multi-angle features
CN114298413A (en) Hydroelectric generating set runout trend prediction method
CN113111096A (en) Abnormity detection method for high-dimensional time sequence working condition data of power generation equipment
George et al. Influence of the sampling interval, decision limit and autocorrelation on the average run length in Cusum charts
Chang et al. Sparse regression for extreme values
CN113688773B (en) Storage tank dome displacement data restoration method and device based on deep learning
Wu et al. Automated Anomaly Detection Assisted by Discrimination Model for Time Series
Han et al. Online aware synapse weighted autoencoder for recovering random missing data in wastewater treatment process

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