CN113723703A - Water quality prediction method and system based on multi-source data fusion and deep learning - Google Patents
Water quality prediction method and system based on multi-source data fusion and deep learning Download PDFInfo
- Publication number
- CN113723703A CN113723703A CN202111044816.0A CN202111044816A CN113723703A CN 113723703 A CN113723703 A CN 113723703A CN 202111044816 A CN202111044816 A CN 202111044816A CN 113723703 A CN113723703 A CN 113723703A
- Authority
- CN
- China
- Prior art keywords
- sub
- data
- parameters
- basin
- water quality
- 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 113
- 238000013135 deep learning Methods 0.000 title claims abstract description 55
- 238000000034 method Methods 0.000 title claims abstract description 46
- 230000004927 fusion Effects 0.000 title claims abstract description 21
- 238000013528 artificial neural network Methods 0.000 claims abstract description 41
- 230000006870 function Effects 0.000 claims abstract description 22
- 238000005192 partition Methods 0.000 claims abstract description 15
- 238000012360 testing method Methods 0.000 claims abstract description 11
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 8
- 238000005457 optimization Methods 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 238000012549 training Methods 0.000 claims description 55
- 238000004364 calculation method Methods 0.000 claims description 23
- 238000009826 distribution Methods 0.000 claims description 15
- 239000010865 sewage Substances 0.000 claims description 9
- IJGRMHOSHXDMSA-UHFFFAOYSA-N Atomic nitrogen Chemical compound N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 claims description 6
- 239000002352 surface water Substances 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 5
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 claims description 4
- 229910052760 oxygen Inorganic materials 0.000 claims description 4
- 239000001301 oxygen Substances 0.000 claims description 4
- 235000004257 Cordia myxa Nutrition 0.000 claims description 3
- 244000157795 Cordia myxa Species 0.000 claims description 3
- PWHULOQIROXLJO-UHFFFAOYSA-N Manganese Chemical compound [Mn] PWHULOQIROXLJO-UHFFFAOYSA-N 0.000 claims description 3
- OAICVXFJPJFONN-UHFFFAOYSA-N Phosphorus Chemical compound [P] OAICVXFJPJFONN-UHFFFAOYSA-N 0.000 claims description 3
- XKMRRTOUMJRJIA-UHFFFAOYSA-N ammonia nh3 Chemical compound N.N XKMRRTOUMJRJIA-UHFFFAOYSA-N 0.000 claims description 3
- 229910052748 manganese Inorganic materials 0.000 claims description 3
- 239000011572 manganese Substances 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 3
- 229910052757 nitrogen Inorganic materials 0.000 claims description 3
- 229910052698 phosphorus Inorganic materials 0.000 claims description 3
- 239000011574 phosphorus Substances 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 2
- 230000008859 change Effects 0.000 abstract description 5
- 238000001704 evaporation Methods 0.000 description 12
- 230000008020 evaporation Effects 0.000 description 12
- 239000011159 matrix material Substances 0.000 description 10
- 230000002354 daily effect Effects 0.000 description 7
- 238000012544 monitoring process Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000001186 cumulative effect Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000007726 management method Methods 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 238000007619 statistical method Methods 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000002310 reflectometry Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Tourism & Hospitality (AREA)
- Data Mining & Analysis (AREA)
- Human Resources & Organizations (AREA)
- Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Artificial Intelligence (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- Entrepreneurship & Innovation (AREA)
- Biomedical Technology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- Game Theory and Decision Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Mathematical Physics (AREA)
- Molecular Biology (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Educational Administration (AREA)
- Primary Health Care (AREA)
- Computing Systems (AREA)
Abstract
The invention provides a water quality prediction method and a system thereof based on multi-source data fusion and deep learning, wherein the prediction method comprises the following steps: determining a basin range and an internal partition thereof; determining basic meteorological parameters of the drainage basin in a prediction time period and preprocessing the basic meteorological parameters; determining sub-basin parameters of the sub-basins in a prediction time period, generating a sub-basin parameter set, and then preprocessing the sub-basin parameter set to generate original data; constructing a full-connection deep learning neural network, and defining a loss function and an iterative optimization algorithm; obtaining a trained deep learning neural network; and inputting a test data set of the original data into the trained deep learning neural network, and predicting to obtain a corresponding water quality parameter. Compared with the prior art, the method and the system thereof do not need to carry out time sequence arrangement on the data, can fuse various types of data, more conveniently carry out water quality change prediction, and improve the efficiency and the practicability of the prediction; the operation is simple and easy to implement, and the method is easy to apply to actual water quality management.
Description
Technical Field
The invention relates to the technical field of water quality prediction, in particular to a water quality prediction method and a water quality prediction system based on multi-source data fusion and deep learning.
Background
The surface water quality prediction is an important support for the environmental management of water bodies in rivers and lakes. Generally, the water quality of rivers and lakes is regularly measured by a water quality monitoring station to obtain water quality data of water bodies, which is used as a basis for water environment treatment. However, the number of water quality monitoring sites is limited, the construction and maintenance costs are high, and it is difficult to achieve coverage of a large spatial range. Therefore, predicting the change in water quality through a mathematical model becomes an economically viable approach.
The water quality prediction is to identify the time variation characteristics of the water quality indexes or the water environment quality conditions of the control units and the response relations of the time variation characteristics to the influence factors such as hydrology, weather and pollution sources in a certain space range. The water quality prediction methods commonly used at present can be mainly divided into two types, one is a numerical simulation method, and the other is a mathematical statistics method. The numerical simulation method is mainly based on the mass conservation law and the mechanical equation, establishes a mathematical model of the generation and migration of water body dirt, and predicts the change of the water quality condition in the water area. Although the method has a physics basis, the calculation is long in time consumption, many parameters required by the calculation are difficult to observe, and the use cost is high. The mathematical statistical method is mainly based on the water quality observation data to establish a statistical correlation model of water quality change, identify the time-space distribution characteristics of the water quality data and predict the change trend of the water quality. The statistical method has high calculation speed, but the required data volume is large, the requirement on the data quality is high, and the difficulty of carrying out multi-factor comprehensive prediction on the water quality in a large space range is still high.
Disclosure of Invention
In order to solve the problems that the water quality prediction in the background technology is single, the prediction accuracy of a prediction method and a model is not high, and the water quality cannot be predicted by adopting various factors, the invention adopts a water quality prediction method and a system thereof based on multi-source data fusion and deep learning. The method and the system adopt meteorological, ecological and socioeconomic multi-source data to predict the water quality in the river basin scale in a large range, and have the advantages of stronger systematicness and practicability, higher prediction speed, larger application scale and wider application range.
In order to achieve the purpose, the invention provides a water quality prediction method based on multi-source data fusion and deep learning, which adopts the following technical scheme:
a water quality prediction method based on multi-source data fusion and deep learning comprises the following steps:
s1, determining a basin range and an internal partition of the basin range, wherein the internal partition comprises a sub-basin partition and an administrative partition;
s2, determining basic meteorological parameters of the drainage basin in the prediction time period;
s3, preprocessing basic meteorological parameters;
s4, determining sub-basin parameters of the sub-basins in the prediction period, and generating a sub-basin parameter set;
s5, preprocessing the sub-basin parameter set to generate original data; dividing the original sample data into a training data set and a testing data set;
s6, constructing a full-connection deep learning neural network, and defining a loss function and an iterative optimization algorithm;
s7, inputting the training data set into a deep learning neural network, and carrying out deep learning neural network training to generate a trained deep learning neural network; and inputting the test data set into the trained deep learning neural network, and predicting to obtain the corresponding water quality parameters.
Further, step S3 specifically includes the following steps:
s31, making a spatial distribution map of a meteorological site and each meteorological data attribute table, wherein the meteorological data attribute table comprises basic meteorological parameters of each time period;
s32, performing spatial interpolation on the meteorological data of each time period of the stations in the range of the whole drainage basin, generating raster data of basic meteorological parameters of each time period of the drainage basin, and storing the raster data into a geospatial database;
and S33, calculating the average value of the basic meteorological parameters of all the grids in the range of each sub-basin, and taking the average value as the meteorological parameters of the sub-basins.
Further, the sub-basin parameters comprise meteorological statistical parameters, ecological parameters, social and economic parameters and water quality parameters; step S4 specifically includes the following steps:
s41, determining weather statistical parameters of the sub-watersheds in the prediction time period, and generating a weather statistical parameter set of the sub-watersheds;
s42, determining ecological parameters of the sub-watersheds in the prediction time period, and generating an ecological parameter set of the sub-watersheds;
s43, determining the socioeconomic parameters of the sub-watersheds in the prediction time period, and generating a socioeconomic parameter set of the sub-watersheds;
and S44, determining the water quality parameters of the sub-watersheds in the prediction time period, and generating a water quality parameter set of the sub-watersheds.
Further, the meteorological statistic parameters of the sub-watersheds comprise the accumulated rainfall of the sub-watersheds j in the previous 7 days before the time t
Cumulative rainfall in the first 14 days
Number of drought days within the first 7 days
And drought days within the first 14 days
Further, step S42 specifically includes the following steps:
s51, acquiring normalized vegetation index raster data of each month in the whole watershed range in a prediction period, and calculating monthly normalized vegetation index data of each sub watershed according to sub watershed partitions;
s52, acquiring annual land utilization data of the whole watershed range in a prediction period, reclassifying the watershed land utilization data according to different types of land to generate grid data of the watershed land utilization, and calculating the utilization data of the different types of land in each sub watershed according to sub watershed areas;
and S53, generating an ecological parameter set of the sub-watersheds.
Further, step S43 specifically includes the following steps:
s431, acquiring socioeconomic parameters of each year and each administrative district in a prediction time period;
s432, converting the socioeconomic parameters of the administrative district in the step S431 into related parameters of the sub-watersheds, namely determining the socioeconomic parameters of each sub-watershed;
s433, generating a social and economic parameter set of the sub-watershed;
the socioeconomic parameters comprise the standing population, GDP, the average population GDP, population density, grain yield and urban sewage treatment rate.
Further, the water quality parameters comprise a dissolved oxygen concentration value, a COD manganese concentration value, an ammonia nitrogen concentration value, a total phosphorus concentration value and a total nitrogen concentration value of the surface water.
Further, step S6 is more specifically: the deep learning neural network adopts three hidden layers; the activation function of each layer adopts a Selu function; a node random discarding mechanism is adopted in the third hidden layer;
the defining loss function and the iterative optimization algorithm are more specifically as follows: the loss function adopts mean square error MSE, the accuracy adopts mean error MAE, and the specific calculation formula is as follows:
wherein z isiIs an actual value, yiIs a predicted value.
Further, the parameters of the deep learning neural network comprise a learning rate, a batch size, a maximum iteration number and a node random discarding rate; the learning rate has a value of 0.001, the batch size has a value of 64, the maximum number of iterations has a value of 1000, and the node random discard rate is 20%.
The invention also provides a water quality prediction system based on multi-source data fusion and deep learning, and the technical scheme is as follows:
a water quality prediction system based on multi-source data fusion and deep learning comprises a multi-source data fusion module, a training sample generation module, a network training module and a prediction module; the multi-source data fusion module is connected with the training sample generation module, and the training sample generation module is connected with the network training module; the network training module is connected with the prediction module; the multi-source data fusion module is used for converting meteorological data of an observation station and social and economic data of an administrative region into sub-basin data; the training sample generation module is used for processing sub-watershed data to generate training samples; inputting the training sample into the network training module, and enabling the training precision of the deep learning neural network to meet the requirement by the network training module through adjusting the parameters of the deep learning neural network; storing parameters such as the weight of the trained deep learning neural network and the like, and generating a network for prediction; and inputting prediction parameters into the prediction module to generate predicted water quality data.
Compared with the prior art, the invention has the advantages and beneficial effects that: the invention adopts meteorological, ecological, socioeconomic multisource data to predict the water quality in the large scale of the basin, and has stronger systematicness and practicability compared with the single water quality prediction in the prior art. In addition, the invention adopts the deep learning neural network to predict the water quality, and has the advantages of higher calculation speed, larger application scale and wider application range. The water quality prediction method and the system thereof have clear theoretical significance, are simple and easy to operate and are easy to apply to actual water quality management.
Drawings
FIG. 1 is a flow chart of the rapid watershed water quality prediction method of the invention;
FIG. 2 is another flow chart of the present invention;
FIG. 3 is a scatter plot of predicted values and measured values of CODMn;
FIG. 4 is a spatial distribution of predicted values of CODMn at a given day.
Detailed Description
The drawings are for illustrative purposes only and are not to be construed as limiting the patent. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The technical solution of the present invention is further described below with reference to fig. 1 to 4 and the embodiment.
A water quality prediction method based on multi-source data fusion and deep learning is disclosed, as shown in fig. 1-2, and comprises the following steps:
s1, determining a basin range and an internal partition thereof;
the internal partitions comprise sub-basin partitions and administrative partitions; the sub-basin subareas are divided by basin digital elevation data through catchment area calculation; the administrative districts are divided according to the administrative district scope of the local level city or the county level city.
In this embodiment, according to the national water resource integrated planning, a south China district is selected as an embodiment watershed. And adopting SRTM 90-meter digital elevation data of NASA to extract digital elevation data of the drainage basin of the embodiment and calculate the catchment area to obtain 138 sub-drainage basins of the south China area. Extracting 75 geographical administrative regions of the national grade city in the south China area by adopting a national basic geographic information center (http:// www.ngcc.cn/ngcc /) distribution diagram (SHP format); and drawing a grade city distribution map of the southern China district.
S2, determining basic meteorological parameters of the drainage basin in the prediction time period;
the basic meteorological parameters comprise
And
wherein i represents a meteorological site, t represents a time period, and a day is taken as a prediction time period;
and
respectively representing rainfall, evaporation and air temperature data observed at meteorological point i at time t. More specifically, the daily value data set (V3.0) of the Chinese ground climate data provided by the Chinese meteorological data network is adopted to extract the daily average rainfall, evaporation and air temperature data of 69 meteorological sites in the embodiment basin from 1 month and 1 day in 2020 to 4 months and 30 days in 2021.
S3, preprocessing the basic meteorological parameters, namely converting the basic meteorological parameters into meteorological parameters of a sub-basin, and specifically comprises the following steps:
s31, making a spatial distribution map of the meteorological site and each meteorological data attribute table: according to the longitude and latitude coordinates of the meteorological sites, ARCGIS software is adopted, the meteorological site space distribution diagram is manufactured according to the longitude and latitude coordinates of the meteorological sites, and rainfall, evaporation and air temperature data of each site and each time period are led into an attribute table of the site distribution diagram.
S32, taking the whole drainage basin as a boundary, more specifically, taking a Chinese south China district as a boundary, and performing spatial interpolation on the meteorological data of each station in each time period by adopting an inverse distance weight method; and (3) making a batch processing program of an inverse distance weight method, generating the grid data of rainfall, evaporation and air temperature of China south China district in each time period (day), and storing the grid data into a geospatial database. Wherein, the rainfall, evaporation and air temperature data are 486 grid graphs each, one graph every day.
S33, making a grid data processing program, extracting rainfall, evaporation and air temperature data of each sub-basin in the south China area in each time period (day), and storing the data into an EXCEL document. One EXCEL document per sub-basin, for a total of 138 EXCEL documents, each containing data for 3 variables (rainfall, evaporation, air temperature) for 486 periods (486 rows, 3 columns). The formula for calculating the average of the data in each EXCEL document, i.e., the rainfall, evaporation and air temperature data of all grids in each sub-basin range, is as follows:
wherein j represents a sub-basin number, t represents a time period, and s is a grid number in the sub-basin j;
representing rainfall, evaporation and air temperature data of the grid s in the range of the sub-basin j;
representing the total number of grids in the range of the sub-basin j;
respectively representing rainfall, evaporation and air temperature data of the sub-basin j at the moment t.
S4, determining sub-basin parameters of the sub-basins in the prediction period, and generating a sub-basin parameter set;
the sub-basin parameters comprise meteorological statistical parameters, ecological parameters, social and economic parameters and water quality parameters. The method comprises the following specific steps:
s41, determining weather statistical parameters of the sub-watersheds and generating a weather statistical parameter set of the sub-watersheds;
and (4) compiling a weather data processing program, importing the EXCEL document in the step (3), and calculating and determining weather statistical parameters of the sub-basin. The meteorological statistical parameters comprise the accumulated rainfall of the sub-basin j in the previous 7 days before the time t
Cumulative rainfall in the first 14 days
Number of drought days within the first 7 days
And drought days within the first 14 days
Where j denotes a sub-basin and t denotes a period. The specific calculation formula of the weather statistical parameters is as follows:
wherein,
counting variable representing drought days, taking rainfall 0.1mm as a drought threshold value, namely when the rainfall is less than 0.1mm on the day (time period t), considering the day (time period t) as drought,
the value is 1.
And then generating a meteorological statistical parameter set of the sub-watersheds:
the parameters of the meteorological statistical parameter set of the sub-watershed comprise: 138 sub-watershed, 486 moments of rainfall, evaporation, air temperature, cumulative rainfall before 7 days, cumulative rainfall before 14 days, days of drought before 7 days, and days of drought before 14 days. The data is stored in EXCEL tables, one EXCEL document per sub-domain, each document containing data for 7 variables for 486 periods (486 rows, 7 columns).
S42, determining ecological parameters of the sub-watersheds in the prediction time period, and generating an ecological parameter set of the sub-watersheds;
s421, acquiring normalized vegetation index (NDVI) raster data of each month in the whole watershed range in a prediction period by adopting satellite remote sensing data, and calculating the monthly NDVI data of each sub watershed according to sub watershed partitions; more specifically, NDVI month data with 1km resolution provided by MODIS satellites are adopted, NDVI data of 16 months in the southern China area from 2020 1 month to 2021 year 4 months are extracted, and 16 raster data graphs are calculated. Compiling an NDVI data processing program, taking 138 sub-watersheds in the south China area as a range, and extracting average data of the NDVI surface of each sub-watersheds and each month; converting the month NDVI data of the sub-basin into day data, wherein the NDVI value of each day is equal to the NDVI of the month of the day. Storing the sub-basin daily NDVI data into the EXCEL document. One EXCEL document per sub-basin, for a total of 138 EXCEL documents, each containing 1 variable (NDVI) data for 486 periods (486 rows, 1 column).
The formula for the calculation of the monthly NDVI data for each sub-basin is as follows:
wherein j represents a sub-basin number, and t representsPeriod (day), m denotes month, s denotes a grid number within the sub-basin j;
representing the NDVI values of the grid s and the mth month in the range of the sub-basin j;
the total number of grids in the range of the sub-basin j is shown;
NDVI of the sub-basin j at the mth month and the tth time is shown; time t is within the mth month.
It is noted that the above NDVI is a remote sensing indicator reflecting the condition of the land cover vegetation, and vegetation is quantified by measuring the difference between near infrared (NIR, vegetation strong reflection) and RED light (RED, vegetation absorption). In addition, NDVI is defined as the quotient of the difference and the sum of the reflectivities of the near infrared channel and the visible light channel, and the value range is always-1 to + 1. Higher NDVI values result when the reflectance of the red channel is lower (greener vegetation) and the reflectance of the near infrared channel is higher. And vice versa.
S422, adopting satellite remote sensing data to obtain annual land utilization data of the whole drainage basin range in a prediction period, and reclassifying the drainage basin land utilization data according to cultivated land, forest land, grassland, water areas, cities, residential land and unused land to generate grid data of drainage basin land utilization. And calculating utilization data of six land types in each sub-river basin according to the sub-river basin subareas. In the present embodiment, the land types include woodland, grassland, water area, city and residential land, and unused land, and the land use data is specifically land use area.
The steps are more specifically; land use year data with 1km resolution ratio provided by Landsat 8 satellites are adopted to extract land use data of China south China district 2020 and 2021 years, and 2 grid data graphs are counted. Compiling a land utilization data processing program, taking 138 sub-watersheds in the south China area as a range, and extracting the area of each sub-watersheds, forest lands, grasslands, water areas, cities and large-amount and unused land for residents; converting the annual land use data of the sub-watershed into day data, wherein the land area of each type of each day is equal to the corresponding land use data of the year of the day. Five types of daily land area data for the sub-watersheds are stored into the EXCEL document. One EXCEL document per sub-basin, for a total of 138 EXCEL documents, each document containing data of 5 variables (five types of land area data) for 486 periods (486 rows, 5 columns).
The calculation formula of the utilization area of the six land types in the invention is as follows:
where j denotes a sub-basin number, t denotes a period (day), y denotes a year, and area denotes an area of one grid within the sub-basin j.
S423, generating an ecological parameter set of the sub-basin according to the utilization areas of different types of land as
The parameters of the sub-basin ecological parameter set comprise: 138 sub-watersheds, 486 NDVI, woodland, grassland, water area, city, and area of the residential and unused land. The data is stored in EXCEL tables, one EXCEL document per sub-domain, each document containing data for 6 variables for 486 periods (486 rows, 6 columns).
S43, determining the socioeconomic parameters of the sub-watersheds in the prediction time period, and generating a socioeconomic parameter set of the sub-watersheds;
the socioeconomic parameters comprise the standing population, GDP, the average population GDP, population density, grain yield and urban sewage treatment rate. The method specifically comprises the following steps:
s431 adopts a Chinese network yearbook database to collect the population of the permanent residence, GDP per capita, population density, grain yield and urban sewage treatment rate of 2020, 2021 and 75 prefectural municipal administration areas within the area of south China, and uses the population density, the grain yield and the urban sewage treatment rate respectively
And
represents; where k denotes an administrative district number within the drainage basin and y denotes a year. The calculation formula of the population density and the population GDP is shown below.
Wherein, AreakIs the area of the kth administrative district.
S432, converting the socioeconomic parameters of the administrative district in the step S431 into related parameters of the sub-watersheds, namely determining the socioeconomic parameters of each sub-watershed;
and converting the socioeconomic data of the administrative region into related parameters of the sub-watershed by adopting an area weighted average method. More specifically, a socioeconomic data processing program is compiled, and socioeconomic data of 75 prefectures are converted into socioeconomic data of 138 subdomains according to the spatial topological relation between the prefecture areas of the prefectures and the subdomains by an area weighted average method and an arithmetic average method. And converting the annual socioeconomic data of the sub-watershed into daily data, wherein the daily socioeconomic data is equal to the corresponding data of the year of the day. The data is stored to the EXCEL document. One EXCEL document per sub-basin, 138 EXCEL documents, each containing 486 periods of 6 variable (6 socio-economic parameters) data (486 rows, 6 columns). The calculation formula of each parameter of the sub-watershed is as follows:
wherein k represents the number of administrative districts in the drainage basin, and N is the total number of administrative districts in the drainage basin; j is the sub-basin number;
the number of the permanent population in the y-th administrative area and the k-th administrative area;
the number of the permanent population of the jth sub-basin in the y year;
the ratio of the area of the kth administrative district in the jth sub-basin to the total area of the kth administrative district is represented;
indicating the area of the kth administrative district inside the jth sub-basin.
The GDP and grain yield were also calculated using the above method. The specific calculation formula is as follows.
Wherein,
the total domestic production value and the grain yield of the y-th sub-basin and the j-th sub-basin are obtained.
And then determining the per-capita GDP, population density and municipal sewage treatment rate of each sub-basin. Since these three parameters belong to the proportional data, the area weighted average method cannot be used to calculate the parameters of the sub-watersheds. And calculating the per capita GDP, population density and urban sewage treatment rate of the sub-drainage basin by constructing a topological relation between the administrative region and the sub-drainage basin and by an arithmetic mean method. The calculation formula of the human-average GDP is shown below.
Wherein k represents the number of administrative districts in the drainage basin, and N is the total number of administrative districts in the drainage basin; j is the sub-basin number;
GDP is the average person in the y-th administrative area and the k-th administrative area;
the GDP is the average GDP of the people in the y-th sub-basin and the j-th sub-basin.
And calculating the population density and the urban sewage treatment rate of each sub-basin by adopting the method of calculating the same GDP, wherein the calculation formula is shown.
S433, generating sets of socioeconomic parameters of the sub-watersheds according to the socioeconomic parameters of the sub-watersheds determined in the step S432 as follows:
wherein t represents a time period (day); y represents year; time t belongs to the y-th year.
S44, determining the water quality parameters of the sub-basin and generating a water quality parameter set of the sub-basin;
the national water quality monitoring data provided by a national surface water quality automatic monitoring real-time data distribution system of a China environmental monitoring central office is adopted to extract the daily average water quality parameters of 204 water quality sites from 1 month and 1 day of 2020 to 4 months and 30 days of 2021 year in China southern China, wherein the water quality parameters comprise the surface waterDissolved oxygen concentration value, COD manganese concentration value, ammonia nitrogen concentration value, total phosphorus concentration value and total nitrogen concentration value are respectively used
And (4) showing. Where t denotes a period (day) and j denotes a sub-watershed.
Because the surface water quality belongs to the parameters of the river water body and does not have the spatial distribution attribute on the surface, the water quality of the sub-basin is calculated without adopting a spatial interpolation method, and the arithmetic mean value of the water quality observation data of the station is directly used as the water quality value of the sub-basin. More specifically, according to longitude and latitude coordinates of water quality sites, an ARCGIS software is adopted to draw a spatial distribution map of the water quality sites, and various water quality index data of each site and each time interval are led into an attribute table of the site distribution map. And (3) adopting an arithmetic mean method, taking the arithmetic mean value of a certain water quality index of all water quality sites in the sub-watershed as the value of the water quality index of the sub-watershed, extracting the values of various water quality indexes of each sub-watershed of China in south China in each time period (day), and storing the values into an EXCEL document. One EXCEL document per sub-basin, for a total of 138 EXCEL documents, each document containing data of 5 variables (5 water quality indicators) for 486 periods (486 rows, 5 columns).
The specific calculation process is shown as the following formula:
wherein,
showing the dissolved oxygen concentration observed in the jth sub-flow field and the qth water quality monitoring station at the time t; m represents the total number of water quality observation stations in the jth sub-flow field.
According to the same method as above, obtaining
Generating a water quality parameter set of the sub-basin according to the water quality parameters, namely
S5, fusing the meteorological statistic parameter set, the ecological parameter set, the social and economic parameter set and the water quality parameter set in the step S4 to generate original data; and then dividing the original sample data into a training data set and a testing data set, wherein 70% of the original sample data is used as the training data set, and 30% of the original sample data is used as the testing data set. All data were then normalized, processed with maximum and minimum normalization:
wherein xmax，xminThe maximum value and the minimum value of the raw data are respectively, x is the raw data, and x' is the normalized data.
More specifically, the meteorological parameters, ecological parameters and socioeconomic parameters of the sub-watersheds are integrated into one data set. The data is stored in EXCEL tables, one EXCEL document per sub-domain, for a total of 138 EXCEL documents. Each document contains 19 variables for 486 periods, including: rainfall, evaporation, air temperature, cumulative rainfall in the first 7 days, cumulative rainfall in the first 14 days, drought days in the first 7 days, drought days in the first 14 days, NDVI, woodland, grassland, water areas, cities and residents, area of unused land, permanent population, GDP, per capita GDP, population density, grain yield and urban sewage treatment rate. The 138 EXCEL document data are then merged into 1 data matrix, the columns of which are 19 variables. The behavior of the matrix is 138 sub-watersheds, 486 days, for a total of 67068 rows. This matrix is designated as TableX and used as an input data matrix for deep learning.
And integrating the water quality parameters of the sub-watersheds into a data set. The water quality data of 138 sub-watersheds are combined into 1 data matrix, and the columns of the matrix are 5 water quality variables. The behavior of the matrix is 138 sub-watersheds, 486 days, for a total of 67068 rows. This matrix is denoted as TableY and is used as the output label matrix for deep learning.
TableX and TableY are normalized and processed with the largest and smallest normalization. Each data in the matrix is subtracted by the minimum value of the column in which the data is located, divided by the difference between the maximum and minimum values of the column. Dividing the matrixes TableX and TableY into a training set and a testing set, and taking 70% of rows as the training set to form TrainX and TrainY; 30% of the rows are taken as a test set, forming TestX and TestY.
S6, constructing a fully-connected deep learning neural network (DNN), and defining a loss function and an iterative optimization algorithm;
constructing a fully-connected deep learning neural network, wherein parameters of the deep learning neural network comprise the number k of hidden layers of the network, the number n of nodes of each hidden layer, an activation function, a loss function (loss), an accuracy calculation function (metrics), a learning rate (learning rate), a batch size (batch size), a maximum iteration number (epochs) and a node random discard ratio (drop). Three hidden layers are adopted, namely, the number of nodes in a first layer is 64, the number of nodes in a second layer is 32, and the number of nodes in a third layer is 16; the activation function of each layer employs a Selu function (scaled exponential linear units). The third hidden layer adopts a node random discarding mechanism, and the discarding rate is 20%. Wherein, the learning rate (learning rate) is 0.001; the size (batch size) takes the value 64 and the maximum number of iterations (epochs) takes the value 1000.
The defining loss function and the iterative optimization algorithm are more specifically as follows: the loss function adopts Mean Squared Error (MSE), the accuracy adopts Mean Absolute Error (MAE), and the training adopts Adam algorithm to optimize the training process. The specific calculation formula is as follows
Wherein z isiIs an actual value, yiIs a predicted value. The smaller the MSE and MAE values are, the closer the predicted value and the true value of the model are, the higher the precision of the model is, and the better the performance is.
S7, taking the training data set as the input of the deep learning neural network, and carrying out deep learning neural network training to generate a trained deep learning neural network; and inputting the test data set into the trained deep learning neural network for prediction to obtain corresponding water quality parameters.
And (3) taking the training data set as the input of the deep learning neural network, taking the water quality parameters of the sub-watersheds as the output labels, and iteratively updating the weight of the deep learning neural network based on the training data until the model converges to generate the trained deep learning neural network. And (5) adopting the trained deep learning neural network to predict the water quality.
And inputting the test data set into the trained deep learning neural network, and predicting to obtain the corresponding water quality parameters. Denormalizing the predicted value to obtain an actual value, drawing a scatter diagram of the predicted actual value and the actually measured water quality value, and calculating the square of the correlation coefficient r (r2)，r2Larger indicates more accurate prediction results. Continuously comparing the predicted value with the measured value, and calculating corresponding MSE and r2According to MSE and r2And adjusting the model parameters, and re-training to ensure the accuracy. Storing the deep learning neural network meeting the precision requirement and the parameter data such as the weight thereof, and providing decision support for the manager.
The denormalization formula is as follows:
y′i＝ymin+yi·(ymax-ymin) (40)
wherein, y'iThe predicted value after reduction; y isiIs a predicted value of the DNN network output. y ismax，yminThe maximum and minimum values of the raw data, respectively.
More specifically, the DNN network is trained using TrainX as an input of the DNN network and TrainY as an output label. The weights of the DNN network are iteratively updated based on the training data until the model converges.
And inputting the TestX into the DNN network, and predicting to obtain the corresponding water quality parameter. And carrying out denormalization on the predicted value to obtain an actual value, and denormalizing the TestY to obtain a water quality measured value. Then, a scatter diagram of the predicted actual value and the measured value is drawn, the square of the correlation coefficient r is calculated (r2), and the larger r2 indicates that the prediction result is more accurate, thereby providing decision support for the administrator. And continuously comparing the predicted value of the surface water quality with the measured value, calculating corresponding MSE and r2, adjusting model parameters according to the values of the MSE and r2, and re-training to ensure the accuracy. Storing the DNN network meeting the precision requirement and the parameter data such as the weight and the like. In the present embodiment, the spatial distribution of the predicted value and the actual measured value of CODMn and the predicted value of CODMn on a certain day are shown in fig. 3 and 4, respectively.
A basin water quality rapid prediction system based on multi-source data fusion and deep learning comprises a multi-source data fusion module, a training sample generation module, a network training module and a prediction module;
the multi-source data fusion module comprises a processing program of an inverse distance weight interpolation program of meteorological data, sub-basin meteorological data, NDVI, land utilization data, socioeconomic data and water quality data, is used for converting the meteorological data of an observation station and the socioeconomic data of an administrative region into the sub-basin data, and is connected with the training sample generation module.
The training sample generation module standardizes the meteorological, ecological, social and economic and water quality data of the sub-watershed and generates a training sample; the training sample generation module is connected with the network training module, and transmits input parameter information of the network model to the network training module, wherein the input parameter information comprises historical meteorological, ecological, social and economic and water quality data.
The network training module adjusts the quantity k of network hidden layers, the number n of nodes of each hidden layer, an activation function, a loss function (loss), an accuracy calculation function (metrics), a learning rate (learning rate), a batch size (batch size), a maximum iteration number (epochs) and the parameters of the deep learning neural network of a node random discarding ratio (drop) until the training precision reaches the requirement, stores the parameters such as the trained neural network weight and generates a network for prediction according to the training precision of the deep learning neural network, and the network training module is connected with the prediction module;
and inputting the meteorological parameters, the ecological parameters, the social and economic parameters and the water quality parameters of a period of time in the future into a prediction module to generate predicted water quality data.
Compared with the prior art, the beneficial effect of this embodiment is: the invention adopts meteorological, ecological, socioeconomic multisource data to predict the water quality in the large scale of the basin, and compared with the single water quality prediction in the prior art, the invention improves the prediction efficiency and the practicability, has clear theoretical significance, is simple and easy to operate, and is easy to apply in the actual water quality management. In addition, the invention adopts the deep learning neural network to predict the water quality, has faster calculation speed, larger application scale and wider application range, overcomes the defects of small calculation range, large calculation amount and low calculation speed caused by depending on numerical simulation in the traditional water quality prediction, and simultaneously avoids the strict requirement on the statistical distribution state of input data when the traditional statistical method is used for predicting the water quality.
It should be noted that the above-mentioned embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions can be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, which should be covered by the claims of the present invention.
Claims (10)
1. A water quality prediction method based on multi-source data fusion and deep learning is characterized in that: the method comprises the following steps:
s1, determining the range of the watershed and the inner area of the watershed;
s2, determining basic meteorological parameters of the drainage basin in the prediction period;
s3, preprocessing basic meteorological parameters;
s4, determining sub-basin parameters in the prediction time period and generating a sub-basin parameter set;
s5, preprocessing the sub-basin parameter set to generate original data; dividing the original sample data into a training data set and a testing data set;
s6, constructing a full-connection deep learning neural network, and defining a loss function and an iterative optimization algorithm;
s7, inputting the training data set into a deep learning neural network for deep learning neural network training to generate a trained deep learning neural network; inputting the test data set into the trained deep learning neural network, and predicting to obtain corresponding water quality parameters;
the internal partition includes a sub-basin partition and an administrative partition.
2. The method of claim 1, wherein: step S3 specifically includes the following steps:
s31, making a spatial distribution map of a meteorological site and each meteorological data attribute table, wherein the meteorological data attribute table comprises basic meteorological parameters of each time period;
s32, performing spatial interpolation on the meteorological data of each time period of the stations in the range of the whole drainage basin, generating raster data of basic meteorological parameters of each time period of the drainage basin, and storing the raster data into a geospatial database;
and S33, calculating the average value of the basic meteorological parameters of all the grids in the range of each sub-basin, and taking the average value as the meteorological parameters of the sub-basins.
3. The method of claim 1, wherein: the sub-basin parameters comprise meteorological statistical parameters, ecological parameters, social and economic parameters and water quality parameters; step S4 specifically includes the following steps:
s41, determining weather statistical parameters of the sub-watersheds in the prediction time period, and generating a weather statistical parameter set of the sub-watersheds;
s42, determining ecological parameters of the sub-watersheds in the prediction time period, and generating an ecological parameter set of the sub-watersheds;
s43, determining the socioeconomic parameters of the sub-watersheds in the prediction time period, and generating a socioeconomic parameter set of the sub-watersheds;
and S44, determining the water quality parameters of the sub-watersheds in the prediction time period, and generating a water quality parameter set of the sub-watersheds.
4. The method of claim 3, wherein: the meteorological statistical parameters of the sub-watersheds comprise the accumulated rainfall of the sub-watersheds j in the first 7 days before the t moment
Cumulative rainfall in the first 14 days
Number of drought days within the first 7 days
And drought days within the first 14 days
5. The method of claim 3, wherein: step S42 specifically includes the following steps:
s51, acquiring normalized vegetation index raster data of each month in the whole watershed range in a prediction period, and calculating monthly normalized vegetation index data of each sub watershed according to sub watershed partitions;
s52, acquiring annual land utilization data of the whole watershed range in a prediction period, reclassifying the watershed land utilization data according to different types of land to generate grid data of the watershed land utilization, and calculating the utilization data of the different types of land in each sub watershed according to sub watershed areas;
and S53, generating an ecological parameter set of the sub-watersheds.
6. The method of claim 3, wherein: the socioeconomic parameters comprise the constant population, GDP, average population GDP, population density, grain yield and urban sewage treatment rate; step S43 specifically includes the following steps:
s431, acquiring socioeconomic parameters of each year and each administrative district in a prediction time period;
s432, converting the socioeconomic parameters of the administrative district in the step S431 into related parameters of the sub-watersheds, namely determining the socioeconomic parameters of each sub-watershed;
and S433, generating a social and economic parameter set of the sub-watershed.
7. The method of claim 3, wherein: step S44 is more specifically: the water quality parameters comprise a dissolved oxygen concentration value, a COD manganese concentration value, an ammonia nitrogen concentration value, a total phosphorus concentration value and a total nitrogen concentration value of surface water.
8. The method of claim 1, wherein: step S6 is more specifically: the deep learning neural network adopts three hidden layers; the activation function of each layer adopts a Selu function; a node random discarding mechanism is adopted in the third hidden layer;
the defining loss function and the iterative optimization algorithm are more specifically as follows: the loss function adopts mean square error MSE, the accuracy adopts mean error MAE, and the specific calculation formula is as follows:
wherein z isiIs an actual value, yiIs a predicted value.
9. The method of claim 8, wherein: the parameters of the deep learning neural network comprise a learning rate, a batch size, a maximum iteration number and a node random discarding rate; the learning rate has a value of 0.001, the batch size has a value of 64, the maximum number of iterations has a value of 1000, and the node random discard rate is 20%.
10. A basin water quality rapid prediction system based on multi-source data fusion and deep learning is characterized in that: the system comprises a multi-source data fusion module, a training sample generation module, a network training module and a prediction module; the multi-source data fusion module is connected with the training sample generation module, and the training sample generation module is connected with the network training module; the network training module is connected with the prediction module;
the multi-source data fusion module is used for converting meteorological data of an observation station and social and economic data of an administrative region into sub-basin data; the training sample generation module is used for processing sub-watershed data to generate training samples; inputting the training sample into the network training module, and enabling the training precision of the deep learning neural network to meet the requirement by the network training module through adjusting the parameters of the deep learning neural network; storing parameters such as the weight of the trained deep learning neural network and the like, and generating a network for prediction; and inputting prediction parameters into the prediction module to generate predicted water quality data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111044816.0A CN113723703A (en) | 2021-09-07 | 2021-09-07 | Water quality prediction method and system based on multi-source data fusion and deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111044816.0A CN113723703A (en) | 2021-09-07 | 2021-09-07 | Water quality prediction method and system based on multi-source data fusion and deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113723703A true CN113723703A (en) | 2021-11-30 |
Family
ID=78682347
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111044816.0A Pending CN113723703A (en) | 2021-09-07 | 2021-09-07 | Water quality prediction method and system based on multi-source data fusion and deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113723703A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114689122A (en) * | 2022-03-31 | 2022-07-01 | 国网北京市电力公司 | Equipment fault monitoring method, device, equipment and medium |
| CN114741461A (en) * | 2022-03-21 | 2022-07-12 | 西安飞蜂智能科技有限公司 | Rural sewage clearing and transportation cost calculation method and system based on topological network |
| CN114936664A (en) * | 2022-03-18 | 2022-08-23 | 桂林电子科技大学 | Watershed non-point source pollution prediction method based on hybrid deep learning model |
| CN115936496A (en) * | 2022-11-29 | 2023-04-07 | 中国环境科学研究院 | Water quality prediction model data treatment standardization method |
| CN117800425A (en) * | 2024-03-01 | 2024-04-02 | 宜宾科全矿泉水有限公司 | Water purifier control method and system based on artificial intelligence |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109242203A (en) * | 2018-09-30 | 2019-01-18 | 中冶华天南京工程技术有限公司 | A kind of water quality prediction of river and water quality impact factors assessment method |
| CN112529234A (en) * | 2019-09-18 | 2021-03-19 | 上海交通大学 | Surface water quality prediction method based on deep learning |
| CN113011992A (en) * | 2021-03-19 | 2021-06-22 | 中国水利水电科学研究院 | River basin agricultural non-point source pollution river entry coefficient measuring and calculating method based on standard data |
-
2021
- 2021-09-07 CN CN202111044816.0A patent/CN113723703A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109242203A (en) * | 2018-09-30 | 2019-01-18 | 中冶华天南京工程技术有限公司 | A kind of water quality prediction of river and water quality impact factors assessment method |
| CN112529234A (en) * | 2019-09-18 | 2021-03-19 | 上海交通大学 | Surface water quality prediction method based on deep learning |
| CN113011992A (en) * | 2021-03-19 | 2021-06-22 | 中国水利水电科学研究院 | River basin agricultural non-point source pollution river entry coefficient measuring and calculating method based on standard data |
Non-Patent Citations (1)
| Title |
|---|
| GUO D.等: "Key factors affecting temporal variability in stream water quality" * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114936664A (en) * | 2022-03-18 | 2022-08-23 | 桂林电子科技大学 | Watershed non-point source pollution prediction method based on hybrid deep learning model |
| CN114741461A (en) * | 2022-03-21 | 2022-07-12 | 西安飞蜂智能科技有限公司 | Rural sewage clearing and transportation cost calculation method and system based on topological network |
| CN114689122A (en) * | 2022-03-31 | 2022-07-01 | 国网北京市电力公司 | Equipment fault monitoring method, device, equipment and medium |
| CN114689122B (en) * | 2022-03-31 | 2023-11-10 | 国网北京市电力公司 | Equipment fault monitoring method, device, equipment and medium |
| CN115936496A (en) * | 2022-11-29 | 2023-04-07 | 中国环境科学研究院 | Water quality prediction model data treatment standardization method |
| CN115936496B (en) * | 2022-11-29 | 2023-09-19 | 中国环境科学研究院 | Water quality prediction model data treatment standardization method |
| CN117800425A (en) * | 2024-03-01 | 2024-04-02 | 宜宾科全矿泉水有限公司 | Water purifier control method and system based on artificial intelligence |
| CN117800425B (en) * | 2024-03-01 | 2024-06-07 | 宜宾科全矿泉水有限公司 | Water purifier control method and system based on artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113723703A (en) | Water quality prediction method and system based on multi-source data fusion and deep learning | |
| CN112765808B (en) | Ecological drought monitoring and evaluating method | |
| CN111159149B (en) | River flow prediction method based on three-dimensional convolutional neural network | |
| Wu et al. | Runoff forecasting using convolutional neural networks and optimized bi-directional long short-term memory | |
| JP2019194424A5 (en) | ||
| CN107292098A (en) | Medium-and Long-Term Runoff Forecasting method based on early stage meteorological factor and data mining technology | |
| Liu et al. | Optimal allocation of water quantity and waste load in the Northwest Pearl River Delta, China | |
| Behmanesh et al. | Estimation of soil temperature using gene expression programming and artificial neural networks in a semiarid region | |
| KR20170005553A (en) | Floods, drought assessment and forecasting techniques development for intelligent service | |
| Zeinolabedini Rezaabad et al. | ANFIS modeling with ICA, BBO, TLBO, and IWO optimization algorithms and sensitivity analysis for predicting daily reference evapotranspiration | |
| CN113139329B (en) | Xinanjiang model parameter calibration method based on hydrological similarity and artificial neural network | |
| Das et al. | FORWARD: a model for forecasting reservoir water dynamics using spatial Bayesian network (SpaBN) | |
| Pal et al. | Hurst exponent approach through rescaled range analysis to study the time series of summer monsoon rainfall over northeast India | |
| Tibangayuka et al. | Evaluating the performance of HBV, HEC-HMS and ANN models in simulating streamflow for a data scarce high-humid tropical catchment in Tanzania | |
| CN111539597A (en) | Gridding drainage basin social and economic drought assessment method | |
| Biazar et al. | Estimation of evaporation from saline-water with more efficient input variables | |
| CN103886220A (en) | Water data discretization method for setting weight based on BP network and Gini coefficient | |
| CN111428936B (en) | River basin rainfall flood availability index measuring and calculating method based on distributed water nodes | |
| Li et al. | A method of rainfall runoff forecasting based on deep convolution neural networks | |
| Zou et al. | Quantitative assessment of adaptive measures on optimal water resources allocation by using reliability, resilience, vulnerability indicators | |
| CN116205136A (en) | Large-scale river basin deep learning flood forecasting method based on runoff lag information | |
| CN118095034B (en) | Hydrologic forecasting method and system based on hydrologic model and LSTM dual coupling under influence of water taking activity | |
| Shakeri et al. | Projection of important climate variables in large cities under the CMIP5–RCP scenarios using SDSM and fuzzy downscaling models | |
| Zhu et al. | Modeling the multiple time scale response of hydrological drought to climate change in the data-scarce inland river basin of Northwest China | |
| CN110110391A (en) | A kind of flood estimation method in the scarce measuring runoff data basin based on region division |
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 | ||
| CB03 | Change of inventor or designer information |
Inventor after: Liu Yueyi Inventor after: Zheng Hang Inventor after: Wan Wenhua Inventor before: Zheng Hang Inventor before: Liu Yueyi Inventor before: Wan Wenhua |
|
| CB03 | Change of inventor or designer information | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20211130 |
|
| RJ01 | Rejection of invention patent application after publication |