CN113516203B - Training method of waterlogging prediction model, prediction method and storage device - Google Patents
Training method of waterlogging prediction model, prediction method and storage device Download PDFInfo
- Publication number
- CN113516203B CN113516203B CN202110914440.8A CN202110914440A CN113516203B CN 113516203 B CN113516203 B CN 113516203B CN 202110914440 A CN202110914440 A CN 202110914440A CN 113516203 B CN113516203 B CN 113516203B
- Authority
- CN
- China
- Prior art keywords
- vector
- training
- prediction
- waterlogging
- pipes
- 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.)
- Active
Links
- 238000012549 training Methods 0.000 title claims abstract description 45
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000003860 storage Methods 0.000 title claims abstract description 13
- 239000013598 vector Substances 0.000 claims abstract description 82
- 238000009825 accumulation Methods 0.000 claims abstract description 19
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 17
- 238000009826 distribution Methods 0.000 claims description 35
- 238000010586 diagram Methods 0.000 claims description 14
- 238000002309 gasification Methods 0.000 claims description 7
- 230000001537 neural effect Effects 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 3
- 238000009394 selective breeding Methods 0.000 abstract description 3
- 238000004590 computer program Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000004088 simulation Methods 0.000 description 4
- 238000012502 risk assessment Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000010792 warming Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Development Economics (AREA)
- Evolutionary Biology (AREA)
- Game Theory and Decision Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application relates to a training method of a waterlogging prediction model, the model, a prediction method and a storage device. The historical data such as rainfall, climate temperature and maximum water accumulation depth which are easy to obtain are used, when in prediction, weather forecast data are combined with pipeline data which are easy to collect on site, an improved gray model is adopted to extrapolate and interpolate characteristic data sequences at different moments and are combined with a long and short time memory network LSTM, and characteristic sequence prediction is carried out. Thereby improving the accuracy of data prediction in the case of small samples. And extracting the feature vector of the big data sample through a plurality of one-dimensional convolution layers. The artificial selection of the features is avoided, so that the whole algorithm can adaptively extract the most representative main component features and automatically weigh the weight ratio of each feature.
Description
Technical Field
The application belongs to the technical field of urban waterlogging prediction, and particularly relates to a training method of a waterlogging prediction model, the model, a prediction method and a storage device.
Background
Various production activities of humans make global climate warming trend serious, and disaster events caused by various extreme climates are frequent. Especially, most of the recent typhoons cause a plurality of days of storm, so that most of the areas are deeply affected by flood disasters and waterlogged ponding, and great inconvenience is caused to the life of people. At present, a plurality of scholars at home and abroad are devoted to a great deal of research aiming at prediction of waterlogging disasters. The content relates to the problems of waterlogging, risk, vulnerability, risk assessment and the like. Research methods can be roughly divided into two groups. One school is knowledge of municipal planning, relying on traditional hydrologic models, hydraulic engineering experience. In urban waterlogging risk assessment based on scene simulation and multi-source data, disaster risk assessment is carried out by combining multi-source data such as hundred-degree thermodynamic diagram, NPP-VIIRS lamplight and the like through a scene simulation method, and modeling is carried out by adopting an HR threshold method; in the literature on-line prediction and early warning system for urban inland inundation based on the Internet of things, a flood drainage and inundation prevention model is constructed on the basis of a regional pipe model. And constructing a Kunming city waterlogging model based on a traditional GIS space analysis model and an SWMM model in a document The Study of Urban Rainstorm Waterlogging Scenario Simulation Based on GIS and SWMM Model. And the characteristics of urban waterlogging and urban effect are researched by designing simulation of different rainfall modes.
In the existing prediction method, characteristic factors are selected, evaluation indexes are artificially defined and selected, noise is difficult to accurately remove in big data, and a main component is extracted. Meanwhile, a large number of data samples need to be collected by adopting a neural network deep learning training model, and the actual engineering needs are hardly met. In the prediction of big data of the waterlogging model, disasters are often closely related to the distribution of the landform features. Thus, the data distribution is not uniform, and the migration of the scene has a great influence on the final result.
Disclosure of Invention
The invention aims to solve the technical problems that: in order to solve the defects in the prior art, a training method for a model for accurately carrying out waterlogging prediction, a waterlogging prediction model, a waterlogging prediction method and a storage device are provided.
The technical scheme adopted for solving the technical problems is as follows:
a training method of a waterlogging prediction model comprises the following steps:
s1, acquiring historical data of rainfall, climate temperature and maximum water accumulation depth in a period of time of a plurality of specific areas;
the rainfall, the gasification temperature and the maximum water accumulation depth are all acquired by data of a meteorological department, the rainfall and the gasification temperature are one-dimensional vectors which are arranged according to time sequence, and the maximum water accumulation depth corresponds to the rainfall and the gasification temperature;
obtaining a pipeline distribution diagram of the specific area, and constructing a pipeline distribution feature vector according to the pipeline distribution diagram, wherein the pipeline distribution feature vector is a one-dimensional vector, each element of the one-dimensional vector is the length and the flow rate of a pipeline flowing into and flowing out of the specific area, the flowing in of the flow rate is a positive value, and the flowing out of the flow rate is a negative value;
the serial rainfall, the climate temperature and the pipeline distribution characteristic vector form an initial characteristic vector A 1 ；
S2, the initial feature vector A is divided into a plurality of one-dimensional convolution layers 1 Extracting features to obtain an extracted feature vector A after extracting features 2 ；
S3, extracting the feature vector A 2 M accumulation cycles are performed through a gray model GM (1, 1) to generate a prediction sequence A 3 。
S4, calculating a prediction sequence A output by the gray model GM (1, 1) 3 With the original sequence A 1 A residual sequence e of (2); e=a 3 -A 1 ；
S5, normalizing the residual sequence e, taking the normalized residual sequence as an input vector for training the long-short-time memory network neural model, and taking the maximum ponding depth as an output vector for training the long-short-time memory network neural model;
s6: and training the long-short-time memory network by using the training input vector and the training output vector to obtain the waterlogging prediction model.
Preferably, in the training method of the waterlogging prediction model, at least 500 pieces of historical data are distributed in each pipeline.
Preferably, in the training method of the waterlogging prediction model of the present invention, there are 4 pipes flowing into and out of the specific area, 3 pipes flowing out, and when the inflow pipe or the outflow pipe is insufficient, the pipes are supplemented with 0, and when the inflow pipe or the outflow pipe is exceeded, the data of those pipes with smaller flow are discarded.
Preferably, in the training method of the waterlogging prediction model, the specific area is a bridge bottom, a tunnel or a low-lying road surface.
The invention also provides a waterlogging prediction model device which is obtained by training the training method of the waterlogging prediction model.
The invention also provides a waterlogging prediction method of the specific area, which comprises the following steps of
Collecting weather forecast data, wherein the weather forecast data comprises rainfall and climate temperature of a certain area, and constructing pipeline distribution feature vectors according to pipeline conditions in the area;
the construction method of the pipeline distribution feature vector comprises the steps that the pipeline distribution feature vector is a one-dimensional vector of a row, each element of the one-dimensional vector is the length and the flow rate of a pipeline flowing into and flowing out of the specific area, wherein the flowing in of the flow rate is positive, and the flowing out of the flow rate is negative;
forming an initial characteristic vector A from rainfall, climate temperature and pipeline distribution characteristic vectors 1 ；
Initial eigenvector A by a multi-layer one-dimensional convolution layer pair 1 Extracting features to obtain an extracted feature vector A after extracting features 2 ；
Will extract feature vector a 2 M accumulation cycles are performed through a gray model GM (1, 1) to generate a prediction sequence A 3 。
Calculating a prediction sequence A output by a gray model GM (1, 1) 3 With the original sequence A 1 A residual sequence e of (2); e=a 3 -A 1 ；
Normalizing the residual sequence e to obtain an input vector;
the input vector is input into the waterlogging prediction model device of claim 5, and the maximum ponding depth is obtained.
Preferably, the waterlogging prediction method for a specific area of the present invention is to supplement with 0 when the inflow pipes or the outflow pipes are not enough for training, and discard the data of those pipes with smaller flow when the inflow pipes or the outflow pipes are out of the inflow pipes for training.
Preferably, according to the waterlogging prediction method for the specific area, when the acquired maximum ponding depth exceeds the early warning value, an alarm is sent out.
The present invention also provides a computer storage medium storing one or more instructions adapted to be loaded by a processor and to perform the above-described method.
The beneficial effects of the invention are as follows:
1. the characteristic sequence prediction is carried out by using historical data such as rainfall, climate temperature and maximum water accumulation depth which are easy to obtain, adopting an improved gray model to extrapolate and interpolate characteristic data sequences at different moments and combining the characteristic data sequences with a long and short time memory network LSTM. Thereby improving the accuracy of data prediction in the case of small samples.
2. And extracting the feature vector of the big data sample through a plurality of one-dimensional convolution layers. The artificial selection of the features is avoided, so that the whole algorithm can adaptively extract the most representative main component features and automatically weigh the weight ratio of each feature.
Drawings
The technical scheme of the application is further described below with reference to the accompanying drawings and examples.
FIG. 1 is a schematic diagram of the content of each step of a training method of a waterlogging prediction model according to an embodiment of the present application;
fig. 2 is a flowchart of a training method of a waterlogging prediction model according to an embodiment of the present application.
Detailed Description
It should be noted that, in the case of no conflict, the embodiments and features in the embodiments may be combined with each other.
The technical solutions of the present application will be described in detail below with reference to the accompanying drawings in combination with embodiments.
Example 1
The embodiment provides a training method of a waterlogging prediction model, as shown in fig. 1, comprising the following steps:
s1, acquiring historical data of rainfall, climate temperature and maximum water accumulation depth in a period of time of a plurality of specific areas; (the maximum water depth corresponds to rainfall and climate temperature data and is from a specific area for a period of time, such as 48 hours)
The rainfall, the gasification temperature and the maximum water accumulation depth are all obtained by data acquisition of a meteorological department and are one-dimensional vectors arranged according to time sequence;
obtaining a pipeline distribution diagram of the specific area, and constructing pipeline distribution feature vectors according to the pipeline distribution diagram, wherein the pipeline distribution feature vectors are one-dimensional vectors in a row, each element of the one-dimensional vectors is the length and the flow rate of a pipeline flowing into and flowing out of the specific area, and the flowing-in flow rate is positive, and the flowing-out flow is negative; typically, the number of pipes flowing into and out of the specific area is 4, the number of pipes flowing out is 3, when the number of the pipes flowing into is less than 4, the pipes are complemented with 0, when the number of the pipes flowing into exceeds 3, the data of the group of pipes with the smallest flow rate is discarded, or similar pipes are combined;
the serial rainfall, the climate temperature and the pipeline distribution characteristic vector form an initial characteristic vector A 1 The method comprises the steps of carrying out a first treatment on the surface of the For example, the inflow pipeline has 3 inflow and 1 outflow, and then (L1, H1, L2, H2, L3, H3, L4, H4, L '1, H'1, L '2, H'2, L '3 and H' 3); wherein Ln is the length of the nth inflow pipeline, hn is the flow of the nth inflow pipeline, L 'n is the length of the nth outflow pipeline, and H' n is the flow of the nth outflow pipeline; wherein the values of L4, H4, L '2, H'2, L '3 and H'3 are 0;
preferably, the historical data under each pipeline distribution is at least 500 pieces or more;
s2, the initial feature vector A is divided into a plurality of one-dimensional convolution layers 1 Extracting features to obtain an extracted feature vector A after extracting features 2 ；
S3, extracting the feature vector A 2 M accumulation cycles are performed through a gray model GM (1, 1) to generate a prediction sequence A 3 . m times finger integer times, predicting sequence A by accumulating loops 3 Stopping when no change occurs.
S4, calculating a prediction sequence A output by the gray model GM (1, 1) 3 With the original sequence A 1 A residual sequence e of (2); e=a 3 -A 1 ；
S5, normalizing the residual sequence e, taking the normalized residual sequence as an input vector for training a long-short-time memory network neural model (LSTM), and taking the maximum ponding depth as an output vector for training the long-short-time memory network neural model (LSTM);
s6: and training the long-short-time memory network by using the training input vector and the training output vector to obtain the waterlogging prediction model.
When the waterlogging prediction model is used, rainfall, climate temperature and pipeline distribution characteristic vectors of weather forecast are required to be input, and finally the rainfall, climate temperature and pipeline distribution characteristic vectors are output by the waterlogging prediction model.
In the pipeline distribution characteristic vector, when the inflow and outflow pipelines are insufficient in the number acquired by training, the quantity is supplemented with 0, and when the inflow and outflow pipelines exceed the number acquired by training, the data of the pipeline group with the minimum flow rate is abandoned.
In particular, the specific area is a bridge floor, a tunnel, a low-lying road surface.
Example 2
The embodiment provides a waterlogging prediction method for a specific area, which comprises the following steps of
Collecting weather forecast data, wherein the weather forecast data comprises rainfall, climate temperature and pipeline distribution feature vectors;
the construction method of the pipeline distribution feature vector comprises the steps that the pipeline distribution feature vector is a one-dimensional vector of a row, each element of the one-dimensional vector is the length and the flow rate of a pipeline flowing into and out of the specific area, wherein the flowing in of the flow rate is positive, the flowing out of the flow rate is negative, when the quantity of the inflow and outflow pipelines is insufficient to be acquired in training, the pipeline distribution feature vector is supplemented with 0, and when the quantity of the inflow and outflow pipelines exceeds the quantity acquired in training, the data of the group of pipelines with the minimum flow rate are discarded;
forming an initial characteristic vector A from rainfall, climate temperature and pipeline distribution characteristic vectors 1 ；
Initial eigenvector A by a multi-layer one-dimensional convolution layer pair 1 Extracting features to obtain an extracted feature vector A after extracting features 2 ；
Will extract feature vector a 2 M accumulation cycles are performed through a gray model GM (1, 1) to generate a prediction sequence A 3 。
Calculating a prediction sequence A output by a gray model GM (1, 1) 3 With the original sequence A 1 A residual sequence e of (2); e=a 3 -A 1 ；
Normalizing the residual sequence e to obtain an input vector;
the input vector was input to the waterlogging prediction model of example 1, and the maximum water depth was obtained.
It should be noted that if a large scale, such as a city-level area region, needs to be calculated, the region needs to be split into nodes, and waterlogging prediction is performed on each node.
When the acquired maximum water accumulation depth exceeds the early warning value, an alarm is sent out, and the water accumulation depth at the position is predicted to exceed the warning water level.
Example 3
The embodiment provides a waterlogging prediction model device, which is obtained by training by using the training method of the waterlogging prediction model in embodiment 1.
Example 4
The present embodiment is a computer storage medium storing one or more instructions adapted to be loaded by a processor and to perform the method of embodiment 1.
Example 5
The present embodiment is a computer storage medium storing one or more instructions adapted to be loaded by a processor and to perform the method of embodiment 2.
In the above embodiment:
1. the characteristic data sequences at different moments are extrapolated and interpolated by adopting an improved gray model and combined with a long and short time memory network LSTM to predict the characteristic sequences by using historical data (weather forecast data combined with pipeline data easy to collect on site during prediction) such as rainfall, weather temperature and maximum water accumulation depth which are easy to obtain. Thereby improving the accuracy of data prediction in the case of small samples.
2. And extracting the feature vector of the big data sample through a plurality of one-dimensional convolution layers. The artificial selection of the features is avoided, so that the whole algorithm can adaptively extract the most representative main component features and automatically weigh the weight ratio of each feature.
With the above-described preferred embodiments according to the present application as a teaching, the related workers can make various changes and modifications without departing from the scope of the technical idea of the present application. The technical scope of the present application is not limited to the contents of the specification, and must be determined according to the scope of claims.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Claims (10)
1. The training method of the waterlogging prediction model is characterized by comprising the following steps of:
s1, acquiring historical data of rainfall, climate temperature and maximum water accumulation depth in a period of time of a plurality of specific areas;
the rainfall, the gasification temperature and the maximum water accumulation depth are all acquired by data of a meteorological department, the rainfall and the gasification temperature are one-dimensional vectors which are arranged according to time sequence, and the maximum water accumulation depth corresponds to the rainfall and the gasification temperature;
obtaining a pipeline distribution diagram of the specific area, and constructing a pipeline distribution feature vector according to the pipeline distribution diagram, wherein the pipeline distribution feature vector is a one-dimensional vector, each element of the one-dimensional vector is the length and the flow rate of a pipeline flowing into and flowing out of the specific area, the flowing in of the flow rate is a positive value, and the flowing out of the flow rate is a negative value;
the serial rainfall, the climate temperature and the pipeline distribution characteristic vector form an initial characteristic vector A 1 ；
S2, the initial feature vector A is divided into a plurality of one-dimensional convolution layers 1 Extracting features to obtain an extracted feature vector A after extracting features 2 ；
S3, extracting the feature vector A 2 M accumulation cycles are performed through a gray model GM (1, 1) to generate a prediction sequence A 3 ；
S4, meterPrediction sequence A output by calculation gray model GM (1, 1) 3 With the original sequence A 1 A residual sequence e of (2); e=a 3 -A 1 ；
S5, normalizing the residual sequence e, taking the normalized residual sequence as an input vector for training the long-short-time memory network neural model, and taking the maximum ponding depth as an output vector for training the long-short-time memory network neural model;
s6: and training the long-short-time memory network by using the training input vector and the training output vector to obtain the waterlogging prediction model.
2. The method of claim 1, wherein the historical data for each pipeline distribution is at least 500.
3. The training method of the waterlogging prediction model according to claim 1 or 2, wherein the number of pipes flowing into and out of the specific area is 4, the number of pipes flowing out is 3, the number of pipes is 0 when the number of pipes flowing into or out of the specific area is insufficient, and the number of pipes having smaller flow is discarded when the number of pipes flowing into or out of the specific area is exceeded.
4. The training method of the waterlogging prediction model according to claim 1 or 2, wherein the specific area is a bridge floor, a tunnel, or a low-lying road surface.
5. A waterlogging prediction model device, characterized in that the waterlogging prediction model device is obtained by training by using the training method of the waterlogging prediction model according to any one of claims 1-4.
6. A waterlogging prediction method for a specific area is characterized by comprising the following steps of
Collecting weather forecast data, wherein the weather forecast data comprises rainfall and climate temperature of a certain area, and constructing pipeline distribution feature vectors according to pipeline conditions in the area;
the construction method of the pipeline distribution feature vector comprises the steps that the pipeline distribution feature vector is a one-dimensional vector of a row, each element of the one-dimensional vector is the length and the flow rate of a pipeline flowing into and flowing out of the specific area, wherein the flowing in of the flow rate is positive, and the flowing out of the flow rate is negative;
forming an initial characteristic vector A from rainfall, climate temperature and pipeline distribution characteristic vectors 1 ；
Initial eigenvector A by a multi-layer one-dimensional convolution layer pair 1 Extracting features to obtain an extracted feature vector A after extracting features 2 ；
Will extract feature vector a 2 M accumulation cycles are performed through a gray model GM (1, 1) to generate a prediction sequence A 3 ；
Calculating a prediction sequence A output by a gray model GM (1, 1) 3 With the original sequence A 1 A residual sequence e of (2); e=a 3 -A 1 ；
Normalizing the residual sequence e to obtain an input vector;
the input vector is input into the waterlogging prediction model device of claim 5, and the maximum ponding depth is obtained.
7. The method according to claim 6, wherein the data of the pipes with smaller flow are discarded when the inflow pipes or the outflow exceeds the number of inflow pipes during training, when the inflow pipes or the outflow is insufficient during training, and the number of inflow pipes is 0-supplemented.
8. The method for predicting waterlogging in a specific area according to claim 6 or 7, wherein an alarm is issued when the obtained maximum water depth exceeds the pre-warning value.
9. A computer storage medium storing one or more instructions adapted to be loaded by a processor and to perform the method of any one of claims 1-4.
10. A computer storage medium storing one or more instructions adapted to be loaded by a processor and to perform the method of any of claims 6-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110914440.8A CN113516203B (en) | 2021-08-10 | 2021-08-10 | Training method of waterlogging prediction model, prediction method and storage device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110914440.8A CN113516203B (en) | 2021-08-10 | 2021-08-10 | Training method of waterlogging prediction model, prediction method and storage device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113516203A CN113516203A (en) | 2021-10-19 |
| CN113516203B true CN113516203B (en) | 2024-02-27 |
Family
ID=78068919
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110914440.8A Active CN113516203B (en) | 2021-08-10 | 2021-08-10 | Training method of waterlogging prediction model, prediction method and storage device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113516203B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI806321B (en) * | 2021-12-28 | 2023-06-21 | 財團法人國家實驗研究院 | Inundation assessment method and computing device |
| CN115186602A (en) * | 2022-06-20 | 2022-10-14 | 郑州大学 | Intelligent decision-making method and system for maintenance of urban underground drainage pipe network |
| CN116469013B (en) * | 2023-06-20 | 2023-09-08 | 云途信息科技(杭州)有限公司 | Road ponding prediction method, device, computer equipment and storage medium |
| CN117332702B (en) * | 2023-11-30 | 2024-03-29 | 南京师范大学 | Waterlogging water depth prediction and multi-factor time sequence analysis method based on ILSTM |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109978235A (en) * | 2019-03-04 | 2019-07-05 | 宁波市气象服务中心 | A kind of flooded water level prediction method of product based on sample learning |
| WO2021082809A1 (en) * | 2019-10-29 | 2021-05-06 | 山东科技大学 | Training optimization method for foreign exchange time series prediction |
-
2021
- 2021-08-10 CN CN202110914440.8A patent/CN113516203B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109978235A (en) * | 2019-03-04 | 2019-07-05 | 宁波市气象服务中心 | A kind of flooded water level prediction method of product based on sample learning |
| WO2021082809A1 (en) * | 2019-10-29 | 2021-05-06 | 山东科技大学 | Training optimization method for foreign exchange time series prediction |
Non-Patent Citations (3)
| Title |
|---|
| PSO-ESN在城市内涝点降雨积水预测中的应用;张梦;赵靓芳;全星;;中国农村水利水电;20190615(第06期);全文 * |
| 基于混合深度神经网络的大气污染预测;宋耀宇;;信息与电脑(理论版);20191225(第24期);全文 * |
| 面雨量在城市内涝预报中的应用试验;李苗;逯张禹;胡志新;;气象与环境科学;20111115(第04期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113516203A (en) | 2021-10-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113516203B (en) | Training method of waterlogging prediction model, prediction method and storage device | |
| Henriksen et al. | Assessment of exploitable groundwater resources of Denmark by use of ensemble resource indicators and a numerical groundwater–surface water model | |
| CN109978235B (en) | Waterlogging water level prediction method based on sample learning | |
| Mitchell et al. | Modelling the urban water cycle | |
| Benedetti et al. | Modelling and monitoring of integrated urban wastewater systems: review on status and perspectives | |
| KR101703981B1 (en) | Control methods of sewage processing facilities using rainfall forecasting | |
| Alamdari et al. | Evaluating the impact of climate change on water quality and quantity in an urban watershed using an ensemble approach | |
| CN113610264A (en) | Refined power grid typhoon flood disaster prediction model | |
| CN111062125B (en) | Hydrological effect evaluation method for sponge type comprehensive pipe gallery | |
| Karyotis et al. | Deep learning for flood forecasting and monitoring in urban environments | |
| Udo et al. | Climate change impacts on runoff regimes at a river basin scale in Central Vietnam | |
| CN116306340A (en) | Method for simulating urban waterlogging risk distribution under different working conditions | |
| Zemzami et al. | Design flood estimation in ungauged catchments and statistical characterization using principal components analysis: application of Gradex method in Upper Moulouya | |
| El Shafie et al. | Radial basis function neural networks for reliably forecasting rainfall | |
| Yihdego et al. | Validation of a model with climatic and flow scenario analysis: case of Lake Burrumbeet in southeastern Australia | |
| Cunha et al. | A decision support model for the optimal siting and sizing of storage units in stormwater drainage systems | |
| Xie et al. | Improving the forecast precision of river stage spatial and temporal distribution using drain pipeline knowledge coupled with BP artificial neural networks: a case study of Panlong River, Kunming, China | |
| CN114943189B (en) | XGboost-based acoustic velocity profile inversion method and system | |
| Sardjono et al. | The application of artificial neural network for flood systems mitigation at jakarta city | |
| CN113239502B (en) | Artificial intelligence image processing-based urban sewage pipe network simulation construction method | |
| Eckersten | Updating rainfall intensity-duration-frequency curves in sweden accounting for the observed increase in rainfall extremes | |
| Zabbah et al. | Prediction of monthly rainfall using artificial neural network mixture approach, Case Study: Torbat-e Heydariyeh | |
| Hadadin | Modeling of rainfall-runoff relationship in semi-arid watershed in the Central Region of Jordan | |
| Kruszyński et al. | Application and digital implementation GIS data to computer modeling of the sanitary sewage network in Podlaskie voivodship | |
| Wang et al. | Study on city rainstorm waterlogging warning system based on historical data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |