Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Water level prediction method based on convolutional neural network

A convolutional neural network and prediction method technology, applied in neural learning methods, biological neural network models, prediction and other directions, can solve problems such as failure to consider, underfitting, and differentiation

Pending Publication Date: 2021-03-23
杭州市水文水资源监测中心
View PDF1 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the water level prediction method described in Comparative Document 1 is only based on the superposition of two existing networks to simply realize the water level prediction function. Specifically, the method in Comparative Document 1 directly inputs the water level sample data into the model, and The main and auxiliary positions of different network layer structure recognition in the neural network are not distinguished, that is, the influence of the weights of different network layers on the prediction results is not considered, the accuracy is poor, and there are disadvantages of underfitting and overfitting

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Water level prediction method based on convolutional neural network
  • Water level prediction method based on convolutional neural network
  • Water level prediction method based on convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0048] Such as figure 1 As shown, the present embodiment provides a water level prediction method based on a convolutional neural network, including:

[0049] S101 Acquiring the target hydrological parameters to be predicted, the target hydrological parameters include one or more of rising water level, upstream water level, downstream water level, discharge flow of upstream reservoir and interval precipitation;

[0050] S102 After standardizing the target hydrological parameters, input them into the trained convolutional neural network, wherein the convolutional neural network includes three layers of convolutional layers and two layers of fully connected layers connected in sequence, and the preset weight of the convolutional layer is greater than Preset weights for fully connected layers;

[0051] S103 predicts the water level value corresponding to the target hydrological parameter within a specified time period after the current time period according to the output of the ...

Embodiment 2

[0074] Such as figure 2 As shown, on the basis of Embodiment 1, in the water level prediction method based on the convolutional neural network in this embodiment, the convolutional neural network is trained based on an online elimination process including:

[0075] S201 Acquire historical hydrological parameters and corresponding historical water level values, as well as real-time hydrological parameters and corresponding real-time water level values;

[0076] S203 Obtain a preset number of reference hydrological parameters according to the real-time hydrological parameters and the historical hydrological parameters after elimination;

[0077] S204 standardizes the hydrological parameters and inputs them into the convolutional neural network to be trained;

[0078] S205 Determine the prediction error percentage according to the prediction value of the convolutional neural network and the corresponding actual water level value, and obtain a trained convolutional neural networ...

Embodiment 3

[0103] On the basis of Example 2, such as Figure 4 As shown, in a specific implementation, this embodiment will refer to the hydrological parameters for standardization processing, and before inputting the convolutional neural network to be trained, it also includes:

[0104] S202 Initialize the weights of the convolutional layer and the fully connected layer to obtain corresponding preset weights;

[0105] Among them, the weights of the convolutional layer are initialized according to the normal distribution random numbers with a mean value of 0 and a standard deviation of 1.

[0106] It should be noted that the convolutional neural network achieves better performance through continuous iterative updating of the weights, and the above iterative updating process is closely related to the weights of the network layer, and bad weights will still lead to gradients in the model training process. Problems such as disappearance and gradient explosion. This scheme initializes the ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a water level prediction method based on a convolutional neural network, and the method comprises the steps: carrying out the standardization of a to-be-predicted target hydrological parameter after obtaining the to-be-predicted target hydrological parameter, and predicting a water level value corresponding to the target hydrological parameter in a specified time period after a current time period according to the output of the convolutional neural network. By combining one or more of the rising water level, the upstream water level, the downstream water level, the upstream reservoir discharge capacity and the interval precipitation capacity and utilizing the powerful feature extraction capacity of the convolutional neural network, the limitation of a traditional water level prediction method is broken through, the precision of water level prediction is greatly improved, the reliability is high, and the method has a wide application range.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, in particular to a water level prediction method based on a convolutional neural network. Background technique [0002] Comparative document 1 (CN201911269292.8) discloses an intelligent water level prediction method based on cyclic neural network and convolutional neural network. Network water level forecasting model for water level forecasting. However, the water level prediction method described in Comparative Document 1 is only based on the superposition of two existing networks to simply realize the water level prediction function. Specifically, the method in Comparative Document 1 directly inputs the water level sample data into the model, and The main and auxiliary positions of different network layer structure recognition in the neural network are not distinguished, that is, the influence of the weights of different network layers on the prediction results is no...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06Q10/04G06Q50/06G06N3/04G06N3/08
CPCG06Q10/04G06Q50/06G06N3/08G06N3/045
Inventor 姬战生章国稳张振林黄薇杨云王英英邱超孟健
Owner 杭州市水文水资源监测中心
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products