CN114757330A - Urban instantaneous water consumption prediction method based on LSTM - Google Patents

Urban instantaneous water consumption prediction method based on LSTM Download PDF

Info

Publication number
CN114757330A
CN114757330A CN202210013577.0A CN202210013577A CN114757330A CN 114757330 A CN114757330 A CN 114757330A CN 202210013577 A CN202210013577 A CN 202210013577A CN 114757330 A CN114757330 A CN 114757330A
Authority
CN
China
Prior art keywords
lstm
water consumption
data
prediction
urban water
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210013577.0A
Other languages
Chinese (zh)
Inventor
赵金伟
罗浩男
雷洲
李爱民
黑新宏
何文娟
杜楠
彭海龙
曹贾隆
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xian University of Technology
Original Assignee
Xian University of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xian University of Technology filed Critical Xian University of Technology
Priority to CN202210013577.0A priority Critical patent/CN114757330A/en
Publication of CN114757330A publication Critical patent/CN114757330A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • 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
    • 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
    • G06Q50/06—Energy or water supply

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Evolutionary Computation (AREA)
  • Economics (AREA)
  • Computational Linguistics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Human Resources & Organizations (AREA)
  • Tourism & Hospitality (AREA)
  • Strategic Management (AREA)
  • Primary Health Care (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses an LSTM-based urban instantaneous water consumption prediction method, which comprises the steps of establishing a prediction mechanism by taking a long-short term memory artificial neural network as a basic regression algorithm, acquiring actual normal operation data of part of urban water consumption units as training samples, learning and training to obtain an LSTM prediction model based on the long-short term memory artificial neural network, and finally obtaining a prediction value output by the LSTM prediction model by taking the urban water consumption in the next time period as the input of prediction. The invention adopts the urban water consumption data sets with similar fluctuation rates to participate in LSTM training, can solve the defect of low accuracy of the historical urban water consumption prediction algorithm, and can well solve the problem of slow convergence of the algorithm.

Description

Urban instantaneous water consumption prediction method based on LSTM
Technical Field
The invention belongs to the technical field of urban water consumption prediction, and particularly relates to an LSTM-based urban instantaneous water consumption prediction method.
Background
With the continuous development of society and economy, the demand of urban water is increasing, but the available water supply is very limited. In order to solve the outstanding contradiction, the local water resource utilization condition must be analyzed and predicted so as to carry out long-term unified planning and management on local water resources and hydraulic engineering. Common urban water consumption prediction methods are divided into three main categories, namely an intuitive prediction method, a time sequence prediction method and a simulation model prediction method. The visual prediction method is a qualitative prediction method, which refers to a method for judging the future water use condition by depending on the visual judgment ability of people, and has certain subjectivity. The time series prediction method only depends on historical observation data to predict future water consumption, and each significant influence factor is not fully considered. The simulation model prediction method can overcome the limitations of the former two prediction methods, such as a prediction modeling method based on a neural network. The method can conveniently and flexibly predict the urban water consumption, thereby having certain practical value.
The factors influencing the urban water consumption are numerous, and a large amount of unpredictability and non-statistics exist, so that the difficulty of water quantity prediction is increased. The prediction of urban water consumption belongs to a nonlinear system, and a neural network shows obvious superiority in the aspect. Most of the current neural networks are trained by using a BP algorithm, and the biggest defects of the neural networks are that the neural networks are easy to fall into local extrema and require long training time.
Disclosure of Invention
The invention aims to provide an LSTM-based urban instantaneous water consumption prediction method, which solves the problems that most of neural networks adopted by existing urban water consumption prediction adopt BP (back propagation) algorithms, are easy to fall into local extreme values and require long training time.
The technical scheme adopted by the invention is as follows: the method for predicting the instantaneous urban water consumption based on the LSTM comprises the steps of firstly establishing a prediction mechanism by taking a long-short term memory artificial neural network as a basic regression algorithm, then collecting actual normal operation data of part of urban water consumption units as training samples, learning and training to obtain an LSTM prediction model based on the long-short term memory artificial neural network, and finally obtaining a predicted value output by the LSTM prediction model by taking the urban water consumption in the next time period as prediction input.
The present invention is also characterized in that,
the method for predicting the urban instantaneous water consumption based on the LSTM comprises the following specific operation steps:
step 1: acquiring urban water consumption data at historical time;
step 2: carrying out normalization processing on the urban water consumption data to obtain a training set;
and step 3: constructing an LSTM prediction model;
and 4, step 4: training the constructed LSTM model according to a training set to obtain an LSTM optimized prediction model of urban water consumption;
and 5: and predicting the urban water consumption at the next moment by the trained LSTM optimization prediction model to generate an urban water consumption prediction result.
The step 1 is as follows:
collecting daily flow data of historical urban water consumption from the hydrological station, wherein the flow data loss rate is required to be less than 10%, otherwise, giving up the daily flow data.
The step 2 is as follows: carrying out maximum value normalization processing on the urban water consumption data, namely dividing each water consumption data by the maximum value;
Figure BDA0003458231900000021
wherein, Fiti(i∈[1,2,…,n]) For the ith sampling data, Fit, in the municipal water consumption datamaxIs the maximum value of the sampled data;
and carrying out normalization processing on the data, wherein the data values of every three continuous time points are a group of training sets, and predicting the data value of the next time point.
And 3, realizing the neural network of the LSTM prediction model by an input layer, a hidden layer and an output layer, wherein the hidden layer has 6 neurons.
The step 4 is specifically as follows:
based on the self-defined parameters in the LSTM prediction model, inputting the training set prepared in the step 2 into the LSTM prediction model, performing forward calculation according to formulas (1) to (6), reducing the loss function value through an LSTM optimizer, and updating the LSTM prediction model and the LSTM optimizer weight parameter omegaf、ωi、ωcAnd ωoObtaining the learned network weight parameters, completing the optimization training of the LSTM prediction model, and obtaining the LSTM optimization prediction model;
the specific process of the forward calculation is as follows:
output from previous time
Figure BDA0003458231900000031
And current time input XtCalculating input gate
Figure BDA0003458231900000032
Forgetting door
Figure BDA0003458231900000033
And last moment memory unit
Figure BDA0003458231900000034
As shown in formulas (1) to (3):
Figure BDA0003458231900000035
Figure BDA0003458231900000036
Figure BDA0003458231900000037
② combine
Figure BDA0003458231900000038
And
Figure BDA0003458231900000039
updating a current memory cell
Figure BDA00034582319000000310
As shown in formula (4):
Figure BDA00034582319000000311
③ pass through the output door
Figure BDA00034582319000000312
Will be provided with
Figure BDA00034582319000000313
To the current time output
Figure BDA00034582319000000314
As in formulae (5) to (6):
Figure BDA00034582319000000315
Figure BDA00034582319000000316
wherein σ (·) and tanh (·) are Sigmoid function and hyperbolic tangent function, respectively; an all indicates a scalar product of two vectors; omegaf、ωi、ωc、ωoRespectively an input gate
Figure BDA00034582319000000317
Forgetting door
Figure BDA00034582319000000318
Last moment memory unit
Figure BDA00034582319000000319
And output gate
Figure BDA00034582319000000320
A weight matrix of (a); bf、bi、bc、boRespectively an input gate
Figure BDA00034582319000000321
Forgetting door
Figure BDA00034582319000000322
Memory unit at last moment
Figure BDA00034582319000000323
And an output gate
Figure BDA00034582319000000324
The offset vector of (2).
The step 5 is as follows:
step 5.1: selecting urban water consumption data of nearly 1 month, and carrying out normalization processing on the data according to the method in the step 2;
step 5.2: and (3) inputting the data subjected to normalization processing in the step 5.1 into an LSTM optimization prediction model to obtain an output value, and multiplying the output value by the maximum value used in normalization in the step 2 to obtain a predicted value of the urban water consumption at the next time point.
The invention is also characterized in that:
the invention has the beneficial effects that:
1. the invention adopts the urban water consumption data set with similar fluctuation rate to participate in the LSTM training, can solve the defect of low accuracy of the historical urban water consumption prediction algorithm, and can well solve the problem of slow algorithm convergence.
2. The invention predicts the urban water consumption by means of the LSTM prediction network, and can automatically realize the prediction of the urban water consumption.
3. The invention uses the LSTM model to realize more accurate convergence result, reduces the original prediction error rate of the urban water consumption to 1.505%, and shows that the model has higher prediction precision and better application value.
Drawings
FIG. 1 is a general flow diagram of the prediction method of the present invention;
FIG. 2 is a graph showing the results of an embodiment of the prediction method of the present invention;
FIG. 3 is a diagram of the prediction of the daily water consumption in a city according to the prediction method of the present invention.
Detailed Description
The invention is described in detail below with reference to the drawings and the detailed description.
The invention discloses an LSTM-based urban instantaneous water consumption prediction method, which is implemented according to the following steps with reference to FIG. 1: the method for predicting the instantaneous urban water consumption based on the LSTM comprises the steps of firstly establishing a prediction mechanism by taking a long-short term memory artificial neural network as a basic regression algorithm, then collecting actual normal operation data of part of urban water consumption units as training samples, learning and training to obtain an LSTM prediction model based on the long-short term memory artificial neural network, and finally obtaining a predicted value output by the LSTM prediction model by taking the urban water consumption in the next time period as prediction input.
The specific operation steps are as follows:
step 1: acquiring urban water consumption data at historical time;
step 2: carrying out normalization processing on the urban water consumption data to obtain a training set;
and step 3: constructing an LSTM prediction model;
and 4, step 4: training the constructed LSTM model according to a training set to obtain an LSTM optimized prediction model of urban water consumption;
And 5: and predicting the urban water consumption at the next moment by the trained LSTM optimization prediction model to generate an urban water consumption prediction result.
The step 1 is as follows:
collecting daily flow data of historical urban water consumption from the hydrological station, wherein the flow data loss rate is required to be less than 10%, otherwise, giving up the daily flow data.
The step 2 is as follows: carrying out maximum normalization processing on the urban water consumption data, namely dividing each water consumption data by the maximum value;
Figure BDA0003458231900000051
wherein, Fiti(i∈[1,2,…,n]) For the ith sampling data, Fit, in the municipal water consumption datamaxIs the maximum value of the sampled data;
and carrying out normalization processing on the data, wherein the data values of every three continuous time points are a group of training sets, and predicting the data value of the next time point.
The LSTM prediction model consists of an input layer, a hidden layer and an output layer; and taking the output of the previous time node as the input of the next time node, and obtaining a predicted value after the final output is subjected to linear regression processing.
Wherein, the neuron of the LSTM model is input by an input gate
Figure BDA0003458231900000052
Forgetting door
Figure BDA0003458231900000053
Output gate
Figure BDA0003458231900000054
The components are connected into a neuron model through a weight and an activation function.
As shown in fig. 2, the working principle can be summarized as follows: the input gate is used for controlling the input access or the access permission of the gate control equipment, and the forgetting gate controls the external state at the last moment
Figure BDA0003458231900000055
The gate control device allows the amount of the gate control device to enter the time t after the time t flows, and the output gate is used for controlling the output value at the time t
Figure BDA0003458231900000056
A door control device that is somewhat visible to the outside.
Firstly forgetting gate outputs by reading last moment
Figure BDA0003458231900000057
And current time input XtOutputs a value between 0 and 1, determines what information we would discard from the neuron, then inputs gates determine that new information is stored in the neuron, and finally outputs gates based on the neuron memory cells
Figure BDA0003458231900000058
Outputting the predicted value
Figure BDA0003458231900000059
The corresponding calculation results are as follows:
output from previous time
Figure BDA00034582319000000510
And current time input XtCalculating input gate
Figure BDA00034582319000000511
Forgetting door
Figure BDA00034582319000000512
And last moment memory unit
Figure BDA00034582319000000513
As shown in formulas (1) to (3):
Figure BDA00034582319000000514
Figure BDA00034582319000000515
Figure BDA0003458231900000061
② combine
Figure BDA0003458231900000062
And
Figure BDA0003458231900000063
updating a current memory cell
Figure BDA0003458231900000064
As shown in formula (4):
Figure BDA0003458231900000065
③ pass through the output door
Figure BDA0003458231900000066
Will be provided with
Figure BDA0003458231900000067
To the current time output
Figure BDA0003458231900000068
As in formulae (5) to (6):
Figure BDA0003458231900000069
Figure BDA00034582319000000610
wherein σ (·) and tanh (·) are Sigmoid function and hyperbolic tangent function, respectively; an all indicates a scalar product of two vectors; omegaf、ωi、ωc、ωoRespectively an input gate
Figure BDA00034582319000000611
Forgetting door
Figure BDA00034582319000000612
Last moment memory unit
Figure BDA00034582319000000613
And output gate
Figure BDA00034582319000000614
A weight matrix of (a); bf、bi、bc、boRespectively an input gate
Figure BDA00034582319000000615
Forgetting door
Figure BDA00034582319000000616
Last moment memory unit
Figure BDA00034582319000000617
And output gate
Figure BDA00034582319000000618
The offset vector of (2).
Examples
The method comprises the following steps: obtaining the water consumption of city
Table 1 shows the recorded data of 12 years of water consumption/hundred million m in 2003-2014 of a city, Shandong province3
Figure BDA00034582319000000622
Figure BDA00034582319000000619
Step two: data normalization processing
From the water usage data in table 1: fitmax＝7.08
The data were normalized and the results are shown in table 2.
TABLE 2 Water consumption normalization data
Figure BDA00034582319000000620
Figure BDA00034582319000000621
Step three: partitioning test sets
Test set partitioning standard: the data values at every three consecutive time points are a set of training sets, and the next time point data value is predicted, with the results shown in table 3.
Table 3 test set partitioning table
Figure BDA0003458231900000071
Step four: construction of LSTM prediction model
Designing an LSTM prediction model and an LSTM optimizer for learning and training; the neural network of the LSTM prediction model is realized by an input layer, a hidden layer and an output layer, wherein the hidden layer has 6 neurons, and the learning rate lr is 0.01. The goal of the LSTM prediction model is to minimize the mean square error between the predicted and true values. Step five: training optimized LSTM prediction model
Based on the self-defined parameters in the LSTM prediction model, inputting the training set prepared in the third step into the LSTM prediction model designed in the fourth step, performing forward calculation according to formulas (1) to (6), reducing the loss function value through an LSTM optimizer, and updating the LSTM prediction model and the LSTM optimizer weight parameter omega f、ωi、ωcAnd ωoAnd obtaining the learned network weight parameters to complete the training of the LSTM prediction model.
Step six: urban water consumption prediction
And (4) taking the last time point in each group of test sets as the input of the LSTM model, applying the LSTM prediction model obtained by training in the fifth step to obtain an output value, multiplying the output value by the maximum value used for normalization to obtain a predicted value of the urban water consumption at the next time point, and obtaining a prediction result shown in a table 4.
TABLE 4 Water consumption prediction based on LSTM model
Figure BDA0003458231900000081
The data results in table 4 show that the relative errors between the predicted values and the actual values in all the years are within 2% except that the difference between the urban water consumption in 2009, 2010 and 2012 is larger and the error is larger in two years before and after the urban water consumption. The difference between the predicted value and the actual value is proved to be small, and the fitting degree is high.
In fig. 3, the dots are real monitoring data of daily water consumption of a certain city (96 data are detected every 15 minutes), and the curve is predicted by using the method of the invention to predict the dots, namely: three consecutive dots are obtained each time, and the next point is predicted to generate a curve. FIG. 3 shows that the prediction result of the method has high precision and is close to actual data.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications and equivalents may be made to the embodiments of the invention without departing from the spirit and scope of the invention, which is to be covered by the claims.

Claims (7)

1. The method for predicting the instantaneous urban water consumption based on the LSTM is characterized by firstly establishing a prediction mechanism by taking a long-short term memory artificial neural network as a basic regression algorithm, then acquiring actual normal operation data of part of urban water consumption units as training samples, learning and training to obtain an LSTM prediction model based on the long-short term memory artificial neural network, and finally obtaining a prediction value output by the LSTM prediction model by taking the urban water consumption in the next period as prediction input.
2. The LSTM-based urban instantaneous water usage prediction method of claim 1,
step 1: acquiring urban water consumption data at historical time;
step 2: carrying out normalization processing on the urban water consumption data to obtain a training set;
and step 3: constructing an LSTM prediction model;
and 4, step 4: training the constructed LSTM model according to a training set to obtain an LSTM optimized prediction model of urban water consumption;
and 5: and predicting the urban water consumption at the next moment by the trained LSTM optimization prediction model to generate an urban water consumption prediction result.
3. The LSTM-based method for predicting instantaneous urban water consumption according to claim 1, wherein step 1 comprises the following steps:
Collecting daily flow data of historical urban water consumption from the hydrological station, wherein the flow data loss rate is required to be less than 10%, otherwise, giving up the daily flow data.
4. The LSTM-based method for predicting instantaneous urban water consumption according to claim 1, wherein step 2 is as follows: carrying out maximum normalization processing on the urban water consumption data, namely dividing each water consumption data by the maximum value;
Figure FDA0003458231890000011
wherein, Fiti(i∈[1,2,…,]) For the ith sampling data, Fit, in the municipal water consumption datamaxIs the maximum value of the sampled data;
and carrying out normalization processing on the data, wherein the data values of every three continuous time points are a group of training sets, and predicting the data value of the next time point.
5. The LSTM-based urban instantaneous water consumption prediction method according to claim 1, wherein the neural network of the LSTM prediction model of step 3 is implemented by an input layer, a hidden layer and an output layer, wherein the hidden layer has 6 neurons.
6. The LSTM-based method for predicting instantaneous urban water consumption according to claim 1, wherein step 4 is as follows:
based on the self-defined parameters in the LSTM prediction model, inputting the training set prepared in the step 2 into the LSTM prediction model, performing forward calculation according to formulas (1) to (6), reducing the loss function value through an LSTM optimizer, and updating the LSTM prediction model and the LSTM optimizer weight parameter omega f、ωi、ωcAnd omegaoObtaining the learned network weight parameters, completing the optimization training of the LSTM prediction model, and obtaining the LSTM optimization prediction model;
the specific process of the forward calculation is as follows:
firstly, the output of the last moment is utilized
Figure FDA0003458231890000021
And current time input XtCalculating input gate
Figure FDA0003458231890000022
Forgetting door
Figure FDA0003458231890000023
And last moment memory unit
Figure FDA0003458231890000024
As shown in formulas (1) to (3):
Figure FDA0003458231890000025
Figure FDA0003458231890000026
Figure FDA0003458231890000027
② combine
Figure FDA0003458231890000028
And
Figure FDA0003458231890000029
updating a current memory cell
Figure FDA00034582318900000210
As shown in formula (4):
Figure FDA00034582318900000211
③ pass through the output door
Figure FDA00034582318900000212
Will be provided with
Figure FDA00034582318900000213
To the current time output
Figure FDA00034582318900000214
As in formulae (5) to (6):
Figure FDA00034582318900000215
Figure FDA00034582318900000216
wherein σ (·) and tanh (·) are Sigmoid function and hyperbolic tangent function, respectively; an all indicates a scalar product of two vectors; omegaf、ωi、ωc、ωoRespectively an input gate
Figure FDA00034582318900000217
Forgetting door
Figure FDA00034582318900000218
Last moment memory unit
Figure FDA00034582318900000219
And output gate
Figure FDA00034582318900000220
A weight matrix of (a); bf、bi、bc、boRespectively an input gate
Figure FDA00034582318900000221
Forgetting door
Figure FDA00034582318900000222
Last moment memory unit
Figure FDA00034582318900000223
And output gate
Figure FDA00034582318900000224
The offset vector of (2).
7. The LSTM-based urban instantaneous water consumption prediction method according to claim 4, wherein step 5 is as follows:
step 5.1: selecting urban water consumption data of nearly 1 month, and carrying out normalization processing on the data according to the method in the step 2;
step 5.2: and (3) inputting the data subjected to normalization processing in the step 5.1 into an LSTM optimization prediction model to obtain an output value, and multiplying the output value by the maximum value used in normalization in the step 2 to obtain a predicted value of the urban water consumption at the next time point.
CN202210013577.0A 2022-01-06 2022-01-06 Urban instantaneous water consumption prediction method based on LSTM Pending CN114757330A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210013577.0A CN114757330A (en) 2022-01-06 2022-01-06 Urban instantaneous water consumption prediction method based on LSTM

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210013577.0A CN114757330A (en) 2022-01-06 2022-01-06 Urban instantaneous water consumption prediction method based on LSTM

Publications (1)

Publication Number Publication Date
CN114757330A true CN114757330A (en) 2022-07-15

Family

ID=82326104

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210013577.0A Pending CN114757330A (en) 2022-01-06 2022-01-06 Urban instantaneous water consumption prediction method based on LSTM

Country Status (1)

Country Link
CN (1) CN114757330A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116451874A (en) * 2023-06-14 2023-07-18 埃睿迪信息技术（北京）有限公司 Urban water consumption prediction method, device and equipment
CN116861192A (en) * 2023-07-27 2023-10-10 广东中山建筑设计院股份有限公司 Urban water consumption prediction method based on SATT-TCN-LSTM model

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116451874A (en) * 2023-06-14 2023-07-18 埃睿迪信息技术（北京）有限公司 Urban water consumption prediction method, device and equipment
CN116451874B (en) * 2023-06-14 2023-09-05 埃睿迪信息技术（北京）有限公司 Urban water consumption prediction method, device and equipment
CN116861192A (en) * 2023-07-27 2023-10-10 广东中山建筑设计院股份有限公司 Urban water consumption prediction method based on SATT-TCN-LSTM model

Similar Documents

Publication Publication Date Title
Ye et al. Predicting electricity consumption in a building using an optimized back-propagation and Levenberg–Marquardt back-propagation neural network: Case study of a shopping mall in China
Zhang et al. Wind speed forecasting based on quantile regression minimal gated memory network and kernel density estimation
Ke et al. Short-term electrical load forecasting method based on stacked auto-encoding and GRU neural network
Zhang et al. Short-time multi-energy load forecasting method based on CNN-Seq2Seq model with attention mechanism
CN110705743B (en) New energy consumption electric quantity prediction method based on long-term and short-term memory neural network
CN108830487A (en) Methods of electric load forecasting based on long neural network in short-term
CN112906982A (en) GNN-LSTM combination-based network flow prediction method
CN105678407A (en) Daily electricity consumption prediction method based on artificial neural network
CN114757330A (en) Urban instantaneous water consumption prediction method based on LSTM
CN107315884A (en) A kind of building energy consumption modeling method based on linear regression
CN111445010B (en) Distribution network voltage trend early warning method based on evidence theory fusion quantum network
CN111292525A (en) Traffic flow prediction method based on neural network
CN111652425A (en) River water quality prediction method based on rough set and long and short term memory network
CN113554466A (en) Short-term power consumption prediction model construction method, prediction method and device
CN112434787A (en) Terminal space energy consumption prediction method based on building total energy consumption, medium and equipment
CN111967666A (en) Comprehensive cooperative scheduling system and scheduling method for river and lake water system
CN113449919B (en) Power consumption prediction method and system based on feature and trend perception
CN114777192B (en) Secondary network heat supply autonomous optimization regulation and control method based on data association and deep learning
CN116167531A (en) Photovoltaic power generation prediction method based on digital twin
Xu et al. Pressure prediction and abnormal working conditions detection of water supply network based on LSTM
CN114936742A (en) Water supply system scheduling agent decision method
CN117422165A (en) Urban water delivery system water quantity prediction method and system based on low carbon emission
CN114429248A (en) Transformer apparent power prediction method
CN114595542A (en) Line loss prediction method
CN116227738B (en) Method and system for predicting traffic interval of power grid customer service

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